WORKPAPERS IN INDONESIAN LANGUAGES, AND CULlURES
Volume 5
UNHAS-SIL
SOUTH SULAWESI SOCIOLINGUISTIC SOIYEYS
1983-1987
THE SUMMER INSTITUTE OF LINGUISTICS
IN COOPERATION WITH
THE DEPARTMENT OF EDUCATION AND CULTURE THE SUMMER INSTITUTE OF LINGUISTICS
IN COOPERATION WITH
THE DEPARTMENT OF EDUCATION AND~ CULTURE
1987 SOUH SULAWESI SOCIOLINGUIST1C SURVEYS 1983-1987
Timothy Friberg, Editor september 1987
TABLE 01' CONTENTS
0. FOREWORD
1. RAMPI AREA (Kabupaten Luwu) Thomas V. Laskowske 1
2. PUSl (Kabupaten Polewali-Mamasa, Northern Section) Timothy Friberg 9
3. PUS2 (Kabupaten Polewali-Mamasa, West-Central Section) Kare J. Stromme 17
4. SEKO AREA (Kabupaten Luwu) Thomas V. Laskowske, Kathryn B. Laskowske 41
5. MANDAR (Kabupaten Polewali Mamasa, Southwestern Section and Kabupaten Majene) Kare J. Stromme, Kari Valkama 59
6. MAMUJU (Kabupaten Mamuju) Kari Valkama 99
7. TORAJA (Kabupatens Pinrang, Enrekang, Tana Toraja, Luwu and Eastern part of Polewali Mamasa) Kari Valkama 119
iii FOREWORD
In 1982-1983Chuck and Barbara Grimes conducted a survey of the languages of South Sulawesi which in its published version (to appear) will certainly stand as the major summary statement of the languages of this province. It is a model to follow both in its literature review and in its data gathering and analysis. Their Languages ofSouth Sulawesi, a definitive result of the cooperative agreement between Hasanuddin University and the Summer Institute of Linguistics, led to a series of further surveys of the province between 1983 and 1987, under that same working agreement, that went both deeper and farther afield than the work that inspired them. Armed with wordlists and questionnaires, members of the UNHAS-SIL cooperative agreement visited every administrative district and subdistrict and multiplied villages in a closer look at the linguistic and sociolinguistic situation of South Sulawesi. Some three hundred wordlists and scores of sociolinguistic questionnaires have considerably expanded our knowledge of the languages of the province. The present volume reports on all but two of those surveys. They are found in chronological order in the following pages. The two not reported on here are the Bugis dialect survey and the Makasar subfamily and southern islands survey. The former is summarized in "A dialect geography of Bugis" (Pacific Linguistics, to appear; the unabridged version is available on microfiche) while the latter will be included in an overview statement of what we know now (or still don't) that was not known in Grimes and Grimes, tentatively entitled "Beyond Grimes and Grimes". The survey reports of this volume are of uneven quality. While some editing has been done, the results presented here reflect a growth in our understanding of the requirements of survey in Sulawesi. 'The language relationships stated in the reports are synChronic rather than genetic, and sociOlinguistic at that. We are very interested in diachronic relationships. We expect that these reports and more specifically the word1ists they are based on will go a long way in determining those relationships. The materials we have gathered are in general available to anyone interested. We welcome individual inquiries even while we seek a vehicle for publishing our word1ists. Based on Grimes and Grimes, these survey reports and the overview statement to follow, we plan further survey at specific locations to answer questions regarding mutual intelligibility and bilingualism. We gratefully acknowledge the assistance in preparing this volume that was willingly given by Barbara Friberg, Sulawesi Program academic affairs coordinator, Anna-Leena Saikkonen, computer coordinator, and Susan Yuanta, administrative assistant.
Timothy Friberg Ujung Pandang September 1987
iv P R A KAT A
Dalam tahun 1982-1983 Chuck dan Barbara Grimes telah melakukan suatu survai bahasa-bahasa di daerah Sulawesi Selatan yang dalam versinya yang telah diterbitkan tentunya akan menjadi ringkasan utama bahasa-bahasa yang ada di propinsi tersebut. Survai in1 merupakan model untuk dicontohi baik dalam tinjauan literatur maupun dalam pengumpulan data dan analisisnya.
Hasil nyata persetujuan kerjasama antara UNHAS dan SIL yang berupa survai bahsa-bahasa daerah Sulawesi Selatan ini telah menyebabkan adanya survai-survai lanjutan di propinsi ini dilakukan antara tahun 1983 dan 1987, yang mana masih di bawah persetujuan kerjasama UNHAS-SIL, yang berjalan lebih jauh dan mendalam dari pada pekerjaan yang telah mereka rencanakan. Dilengkapi dengan daftar-daftar kata dan kuesioner, para anggota kerjasama UNHAS-SIL mengunjungi daerah kabupaten dan kecamatan serta beberapa desa untuk melihat dari dekat keadaan linguistik dan sosiolinguistik di daerah Sulawesi Selatan. Kurang lebih 300 daftar kata dan angka-angka dari kuesioner sosiolinguistik ini benar-benar telah menambah pengetahuan para peneliti tentang bahasa-bahasa di propinsi ini.
Volume ini menyajikan semua laporan, kecuali dua survai dari survai-survai terakhir. Laporan ini disusun secara kronologis pada halaman-halaman berikut. Kedua survai yang tidak dilaporkan di sini adalah survai dialek bahasa Bugis dan survai subfamili bahasa Makasar serta pUlau-pulau di bagian selatan Sulawesi. Ringkasan survai bahasa Bugis dalam "A Dialect Geography of Bugis" (akan terbit dalam Pacific~ Linguistics; versi yang lengkap tersediapada mikrofis), sedanghkan survai subfamili bahasa Makasar akan dimasukkan dalam suatu ikhtisar tentang hal-hal yang sudah ataupun belum diketahui dan juga yang belum ada dalam tulisan Grimes and Grimes yang berjudul "Beyond Grimes and Grimes".
Laporan survai dalam volume ini belum begitu lengkap. Sementara beberapa pemeriksaan sudah dilakukan, hasil yang dikemukakan di sini menggambarkan suatu perkembangan dalam pemahaman peneliti tentang syarat-syarat survai di daerah Sulawesi.
Hubungan-hubungan bahasa yang disampaikan dalam laporan ini lebih berbentuk sinkronis dari pada genetik dan sosiolinguistik. Kami sangat tertarik dengan hubungan-hubungan diakronis. Kami mengharap agar laporan ini
v dan terlebih yang berdasarkan pada daftar-daftar kata akan membantu ~alam menentukan hUbungan-hubUngan tersebut. Bahan-bahan yang telah kami kumpulkan tersedia bagi mereka (siapa saja) yang tertarik. Sementara kami meneari sarana untuk menerbitkan daftar-daftar kami, kami menyambut permintaan perorangan.
Berdasarkan tulisan Grimes and Grimes, laporan-laporan survai dan ikhtisar ini akan menyusul, kami mereneanakan survai lanjutan pada daerah-daerah tertentu untuk menjawab pertanyaan-pertanyaan mengenai inteligibilitas dan dwibahasa.
Kami mengueap terima kasih atas kerelaan ibu Barbara Friberg (Koordinator Bidang Akademik Program Sulawesi), nona Anna-Leena Saikkonen (koordinator Bidang Komputer) dan nona Susan Yuanta (Asisten Administrator) yang telah membantu~ dalam persiapan-persiapan volume ini.
Timothy Friberg Ujung Pandang September 1987
vi UNHAS-SILSociolinguistic Survey: Rampi Area (Kabupaten Luwu) Thomas V. Laskowske
INTRODUCTION
The survey team made their visit to Rampi 15-24 November 1983. We visited each village presently occupied and took wordlists. In several villages sociolinguistic information was also gathered. We walked the trails from Leboni as far as Tedeboe (see map). We were told that the area from Leboni to Rato is at present abandoned and therefore we did not try to visit it. The team participants were: Francis B. Dawson, M.A. - Survey Leader Donald Barr, M.A. Eui Jung Kim, M.A. Thomas V. Laskowske, M.S. Mr. Barr came~ from Palu, Sulawesi Tengah. The other three participants came from Ujung Pandang.
InNEIARY
15 November - Drove from Ujung Pandang to Palopo. Driving time 7 1/2 hours. 16 " - Met Bupati Drs Tawakkal and obtained letters to the camats of Masamba and Lembong. Drove 1 hour to Masamba. 17 " - Met with Camat H.M. Djunaid of Masamba and obtained letters to the kepala desas. Cheeked in with the police and visited the P dan K office. 18 " - The MAF plane (with Mr.~ Barr~ aboard) flew us to Onondoa, Rampi. Stayed that day and night in Onondoa. Started gathering information. 19 " - Split into two teams. Mr. Dawson and Mr. Kim (Team I) went north. Stayed overnight in Bangko. Mr. Barr and Mr. Laskowske (Team II) went south. Stayed overnight in Leboni. 20 " - Team I went to Tedeboe and spent the night. Team II returned to Onondoa. They also stopped in Su1aku to witness the installation of the new kepala kampung. 21 " - Teams I and II met in Dodolo to expedite the survey there. 22 " - Teams stayed in Onondoa evaluating the information and talking to more people. 23 " - Left Rampi for Masamba via MAF plane. Drove to Makale (3 hours) and overnighted there. 24 " - Returned to Ujung Pandang, 7 1/2 hours driving time.
RAMP I 1 Map of Rampi Area
RAMP I DESCRIPTION OF DAERAH RAMPI
Geography (see map)
Rampi lies mainly in an upland valley at an altitude of about 1000 meters. The five villages of Leboni, Sulaku, Onondoa, Meloi, and Dodolo are near the main river of Koro, which eventually becomes the Lariang. This valley is about 15 km long. The villages of Bangko and Tedeboe are separated from Dodolo and the rest of Rampi by a mountain, whose height is perhaps 400 meters above Dodolo. Bangko's elevation is higher than that of the other Rampi villages. Its nearby streams are smaller than the rivers near the other villages. Over another, smaller mountain from Bangko, Tedeboe is located on a tributary of the Koro river, joining the Koro in Central Sulawesi. The nearest group outside of Rampi is Bada, one day's walk to the north of Dodolo in Central Sulawesi. To get to Rampi from Masamba requires a 3-day walk, 81 km. The soil is sandy clay. There is plenty of sand in the rivers. The parent material is granite. Many parts of the valley which are not cultivated are open grassland rather than jungle.
Government
Three desas comprise Rampi. Desa Leboni - Leboni Sulaku Desa Onondoa - Onondoa Meloi Desa Rampi - Dodolo Bangko Tedeboe
The kepala desas are not permanent residents of Rampi. However, the kepala desas were in Rampi to welcome the camat's pending visit. Each village has hansips, village policemen, and in each case they are young men. Each kepala desa also has a local wakil and/or a tata usaha. All were very helpful.
Recent His tory
Kahar Muzakkar's forces controlled the Rampi area during the Rebellion. In 1952 and 1953 most of the Rampi people fled to Central Sulawesi. Subsequently the Rampi people lived outside of their homeland for some fourteen years. They are now moving back,
RAMP I 3 beginning in 1966 after the Indonesian army put down the Rebellion. We are told that there were 5,000-7,000 people living in Rampi. Now, however, there are under 1,500.
Population of Rampi
Village Population* Houses (by our count)
Leboni 100 17 Sulaku 200 29 Onondoa-Heloi almost 500 60 Dodolo 160 19 Bangko 130-180 23 Tedeboe 220-250 36
*Information gathered from Masamba police and Rampi village leaders
There are 10 villages of Rampi people scattered in Central Sulawesi. It was said that some are happier there and have a good life. Others wish to return, but are still afraid.
Culture
Houses Houses are made from wood and bamboo. Many have sawn lumber as part of the construction. Most roofs are thatched, but some have wood shingles or corrugated iron. Every house and garden has a surrounding fence to keep out the cattle. Houses are on stilts about 1 1/2-2 meters off the ground. A small front room is used for entertaining. A bedroom is on one side and kitchen in the back. Food The main food is rice, which is stored unthreshed. Rice is threshed and pounded for a few days' supply at a time. Their vegetables are mostly greens, although there are also some corn· and cassava in the gardens. Their gardens are quite small. There is an abundance of meat including water buffalo, beef, chicken, deer, wild pig and some fish. Clothing Some of the older women still wear the traditional two-layered dresses. One old man we saw still wore the traditional head cloth/turban. Otherwise clothing is modern. Men leaving the village often take their bush knives with them strapped around their waists. Manner People seemed reserved, although they were willing to talk about themselves. Men never greeted women on the street and women seemed quite shy. The people in Bangko and Tedeboe were not as sophisticated as those in the other villages.
4 RAMP I Music We observed at least two bamboo flute bands of school children. Many youths in Bangko and Tedeboe played ukeleles. In Tedeboe they performed traditional dances.
Bcona.y
There is plenty of food, so the Rampi people only import sugar, salt and a few specialties. They sell rice and produce in Bada, but only take dried meat ·or coffee to Masamba, because of the carrying weight. They also buy soap, cloth, and kerosene in Masamba, but don't buy anything in Bada, because the prices are high there. Traders often come on foot or by horse. Cash income largely comes from selling cattle. They also claimed to have lots of coffee trees, but we did not see enough evidence to affirm whether it is an important part of their economy. We observed a number of sewing machines. Also, quite a few men seemed to be skilled in carpentry, sawing boards and making furniture. Tools consisted of planes, saws, hammers, and chisels. Transportation and communication was their biggest felt need for the economy. There are high hopes for a road to be built from Masamba. There are no stores or markets in Rampi due to its isolation.
Health
People are well-fed and seem pretty healthy. There is no Pus-Kes-Mas ~n Rampi. Although there is a building for one in Qnondoa, no one has been assigned there. A~anteri in Sulaku gets medical supplies twice a year, but they are soon used up, and people lack even basic medicines. So, when people do have medical problems, they have a very difficult time. Eye infections, goiter, emphysema, skin disease, dysentary, malaria, and tuberculosis are known to occur in the area.
Education
There are Seko1ah Dasars at Sulaku, Qoondoa, and Dodolo, each with 2 teachers. Of the 6 teachers, all but one are from Rampi. Virtually all the children in those villages attend school. A few boys from Leboni walk more than 1 hour to go to school in Sulaku. Also, about 10 make the 3-hour trip to Dodo10 from Bangko. Even though Bangko and Tedeboe have school buildings, there are no paid teachers available. Tedeboe does have a volunteer teacher who holds classes when he can. Those two villages very much want their schools to have regular classes. The Dutch started schools in 1918 and 1923 in Rampi, so there are many adults who attended Sekolah Dasar and speak Bahasa
RAMP I 5 Indonesia. Yet, the teacher in Dodolo said that a number of parents are reluctant to buy school supplies for their children or to encourage them to go beyond Sekolah Dasar.
Religion
The area is basically Christian. There is a Gereja Kristen Sulawesi Tengah (GKST) church in every village. There is also a Pentacostal group in Tedeboe meeting in private housing. The Islam population of Rampi centers at Sulaku. Today, half the village of Sulaku is Islam and there are some Muslims living in Onondoa. Two local pastors expressed a need for written materials to teach their congregations. Church attendance seemed to vary. Only about 20 attended at Leboni, but more than 100 attended at Tedeboe. Services were held in Bahasa Indonesia, but afterwards announcements were made in Bahasa Rampi.
Language
The people we met, for the most part, spoke to us in Bahasa Indonesia. However, they converse with one another in Bahasa Rampi. There was some evidence in Tedeboe and Bangko that some of the older people did not use Bahasa Indonesia. A study should be made to determine the extent of bilingualism. We project that it would not be as extensive as it appears on the surface, because most of the peopl& we talked with were the better educated leaders in the community. The group as a whole, however, has probably needed to become more fluent in Bahasa Indonesia beca~se of their displacement and necessity of communicating with people in other language groups. We are told the Rampi people residing outside of the Rampi area have also learned the local languages where they are living. In spite of this, they have retained their mother tongue.
MUtual Intelligibility
Within Rampi itself there is some dialect variation, but we would still consider it one language with mutual intelligibility throughout. Comparing three villages showed cognate counts greater than 90%.
6 RAMP I Leboni
96 Dodolo
91 93 Tedeboe
% Cognate
Probably because of the mountain barrier between Tedeboe and Dodolo, it appears that the dialects of Bangko and Tedeboe diverge the most from the rest of the villages. Some of the differences:
Indonesian Leboni Dodolo Tedeboe
adik [hidio'] [tua~i] [hint) duri [ruil [ruil [t~m' tohu)
Prestige Dialect
Rampi speakers consider Bangko to be the place of cultural and linguistic origin, an idea supported by loeal legend. Therefore, it seems wise that the Bangko dialect should have a strong place in any proposed program. The Leboni dialect was referred to as less Rampi because the people there speak "fast and strong".
IntegratiOD into Indonesian Society
The people are eager for a road to help their economy and for schools. So, they seem open to outside help and involvement in the larger national society. An airstrip was opened in August/September, 1983.
COBCLUSIOB
The survey found that the people of Rampi are a single language group. Wordlists show a cognate word count of over 90% similarity among the seven villages visited. A Bada word count (to the north) is only 62% cognate (Salombe, Barr and Barr, 1979). Seko to the west is 38% cognate (Grimes and Grimes, to appear). A single language program should meet the UNHAS-SIL goals for the area. Visits to the Rampi villages outside of the Rampi area
RAMP I 7 will be necessary to ascertain the extent of the program we should undertake.
References
Gr~es, Charles and Barbara Grimes. Languages of South Sulawesi, to appear.
Salombe, C., Don Barr and Sharon Barr, 1979. Languages of Central Sulawesi, Hasanuddin University, Ujung Pandang.
8 ~I URBAS-SIL Sociolincaiatic Survey: Kabupaten Polewali-Mamasa, Northern Section Timothy Friberg
The survey was conducted by Eui Jung Kim and T~othy Friberg, December 6-16, 1983. We left Ujung Pandang the morning of December 6 by Daihatsu jeep and spent the night in Polevali. The bupati's office was contacted to arrange an appointment for the following morning. December 7. We were warmly received at the bupati's office by a clerk and other staff. After some discussion there, we were taken to see the bupati. It was a cordial meeting. We were given a map of the kabupaten and letters of introduction to the five relevant kecamatans of the nine within kabupaten Po~as. Then we paid a visit to the Department of Education and Culture local office. From that meeting we left to begin the drive up the road to Mamas a , going only as far as Malabo, where we spent the night in the home of the camat of kecamatan Mamasa. December 8. Having left the jeep in Malabo, we headed west by foot for Mambi with one porter-guide. Distance 28 kilometers, travel time 9 hours. In Mambi we met and stayed with the camat for the night. December 9. We spent the day in Mambi, getting our presence squared away with the police, visiting various local people and the one foreigner in town, Ken George, and generally getting a feel for the area. Again we spent the night with the camat. December 10. We left Mambi for Aralle with one porter-guide. Distance 11 kilometers, travel time 4 hours. We got a wordlist and filled out a questionnaire. We spent the night with the kepala desa. We discussed a problem with him of our itinerary not taking us to the next kepala desa's location (desa Bumal) before we had already spent several nights in his area of jurisdiction. The kepala desa of Aralle relieved us by telling us that our intended path would take us the next day PUSI 9 The porter went ahead too. We arrived in Tampa1opo after S1X hours, forty minutes on the trail; distance twelve kilometers. We stayed at the house of the guru iemaat. In addition to wordlist and questionaire, we participated in the first of three funeral services for a small boy who died just previous to our arrival in the village. December 13. We left Tampalopo for Tabulahan. Travel time was six hours, distance twelve kilometers. Enroute at Periangan we took a wordlist and filled out a questionnaire for the Makki language (Kalumpang). In Tabulahan we stayed the night with the kepala kampung, the kepala desa being away. We took a word list and filled out a questionnaire in Tabu1ahan. December 14. We left Tabu1ahan for Sodangan with one porter-guide, stopping to rest at Taora enroute. We stayed the 'night in Sodangan with a husband-wife teaching team. We did not take a wordlist or questionnaire, but asked many questions. Distance 21 kilometers, time nine hours, forty minutes. December 15. We left Sodangan for Tapalina with a porter-guide. Enroute we changed plans and pushed straight through to Mambi. Travel distance, 22 kilometers, nine hours travel time. In Mambi we stayed with the camat again, spoke again with Ken George and secured horses for the next day. December 16. We left Mambi for Malabo by horseback in the company of the kepala of desa Bumal and the horae owners. In Malabo we made contact again with the camat's family and then drove down the mountain to Polewali. In Polewali we had light refreshments, and drove on to Ujung Pandang, arriving at midnight. Trail conditions in our great circuit were poor to good. We were rained on three times of no more than thirty minutes total. In a number of places the trail was pure mud, caused by the rainy season, but maintained by the heavy jungle overgrowth. Rockiness of the trails was in many places more grievous than the mud. Survey objectives and results One of the survey objectives was to ascertain whether the dialect situation for the PUS language area was the same as the Grimeses wrote about in their Languages of South Sulawesi (to appear). Specifically, whether the two dialects for which they had no wordlists were indeed within the PUS language or whether they existed at all. In particular, we sought information ab~ut dialects named by the Grimeses as Ulunda and Tapango. In the latter case we were not able to confirm its existence directly, for our survey stopped short of its goal of surveying the entire PUS area; that is, we did not visit the southern reaches of the language area. However, we were told by several people that Tapango is indeed a dialect of PUS. Its linguistic relationship to the rest of the language is yet to be determined. In the 10 PUSl former case, our numerous inquiries as to the location and sta~us of U1unda were all without affirmative response. Summarizing the answers of many, U1unda is thought to be either U1u Mandar or U1u Ma1unda (U1u[ma1u]nda), a dialect of Mandar. Those with knowledge of the villages west of Mambi town said that the few near ones were all the same dialect as that spoken in Mambi. Farther west there is open jungle before the coastal Mandar dialects are reached. Ignoring Tapango and discounting Ulunda, we find the linguistic relationship within PUS to be close to the following: (Buma1) Tabu1ahan Ara11e Bambang I (Mambi)' I Rantebu1ah.!!r Mehala'an Aralle and Tabu1ahan, though separated by intrusive Bambang, are close dialects (89%) and form one related set. Bambang and its extension into desa Buma1 are popularly considered to be another unified set, though we got no word1ist from desa Bambang. (Sa1udadeko and Tampa1opo, both in desa Buma1, relate at 94%). Rantebu1ahan and Meha1a'an, though we didn't test them, appear to be another related set, both in the people's perceptions and perhaps due to their sharing adjacent tracts of land. Mambi seems to be without its own clear identify as a dialect of PUS. This perhaps stems from its being adjacent to all three of the other sets, Ara1le(-Tabu1ahan), Bambang and Rantebu1ahan-Meha1a'an. The closest relationship, if one is to be strongest, is probably Mambi with Bambang. Further, Mambi is the center of the kecamatan in perhaps every sense but linguistic and the Christianity aspect of religion (it is the Muslim center). As such there is a mix of p~op1e in its population from allover the PUS area as well as from without. Word1ists taken from Mambi would have to ensure PUSl 11 that not only were the informants born in desa Mambi, but also their parents. As shown on the following matrix, the cognate percentage betweeen Ara11e-Tabu1ahan and Bambang (Saludadeko and Tampalopo) is low enough (74%) to cause serious doubt as to whether they can be considered dialects of a single language. The view that we tentatively suggest, as propounded by field scholar Ken George, is that perhaps the dialects or languages traditionally grouped as PUS were already distinct languages that because of outside pressures grouped together in a defensive arrangement as a policy of convenience and survival. This alliance later came to be viewed as a cultural and linguistic entity. This theory needs to be examined in light of extensive linguistic and cultural research in the area. Periangan -74 Tampalopo 72 94 Saludadeko 66 73 74 Aralle 64 74 74 89 Tabulahan The Makki dialect (kampung Periangan) noted in the matrix is a member of the Kalumpang language to the north and northeast of PUS. Some thirteen villages of Makki speakers have moved into the northern reaches of PUS in the last couple generations. Locally produced books In our survey travels we came across two books of interest to an understanding of PUS. The first is Mamasa a historical, cultural and tourist commentary on the Polewali-Mamasa district written by Arianus Mandadung. The second is Ungkapan SejaTah ~ Budaxaan di kabupaten Polewali Mamasa Sulsel by Abdul Azis Samar and Arianus Mandadung. Note that both books share·one author. IleligioD PUS is generally divided between adherents of Islam and Christianity with some few thousands following traditional animism, called Maplondo locally. Desas Mambi and Aralle are predominantly Islam. Bumal, Tabulahan and Bambang are Christian. 12 PUSl The churches are of GTM CGereja Toraja Mamasa) affiliation and they look eastward to Mamasa for direction. Education Most villages have their own primary school, though in the more remote villages there are no schools or only schools shared among several villages. There is an SMP in Mambi as well as a private SMP in Aralle. Those finishing SMP have to leave the kecamatan for SMA schooling. We were told at various times that these students attended schools in Mamuju, Mamas a and Ujung Pandang. There are Muslim schools in many Muslim towns. There is general interest in education, but we made no systematic effort to look at it closely. Economy Most villages if not all villages, then desas - are independent in the basics of food and housing. The family unit owns or farms either sawah or ladang for rice and other food needs. We noticed that fresh vegetables were scarce, though cassava leaves were in abundance. Some corn was evident. Fish, especially dried fish, seemed to be the basic meat product. Very few domestic an~als, apart from some chickens (and dogs), were in evidence. The people eat what they raise. Only coffee is widely sold as a cash crop. In exchange for coffee, people buy clothing, some food supplements and basic tools. There is a revolving market operating in six locations. Sunday there is no market. Traveling merchants buy and sell at the markets on this weekly cycle. Pack horses carry their wares. Geography The area of kecamatan Mambi is basically mountainous. A number of rivers drain the area, either generally flowing south or west to the sea. Indeed pitu U1unna Salu means 'seven river heads'. These rivers are too shallow, rocky and fast flowing to be used for transportation within the PUS area. Although many hectares of forest have been felled for ladang and developed into sawah through the generations, there are still vast tracks of virgin forest. The land is rugged enough to make level areas for airstrips a rare commodity. Trails between villages generally follow the lay of the land, so travel distances are a great deal longer than air distances. PUSl 13 Health aad lledicine The people often complained of poor health and their remoteness from healthcare service. 'Poor health' probably means occasional sickness, for we saw no signs of endemic sickness nor heard complaints of malaria. There is a clinic in Mambi and a smaller one in Aralle. North of Aralle there is nothing available. Bilingualism In every village we visited we found speakers of Indonesian, regardless of how remote we were. But in all cases we were talking to village leaders, so we don't know the Qepth nor breadth of bilingualism within village life. Transportation and communication The regular Indonesian governmental hierarchy is known and appreciated right down to the village level. Kepala desas frequently travel to Mambi for government meetings. Villagers travel at least as far as the desa capital for marketing. We frequently noted that men also travel outside the area. Such travel approximates river flow: in the northern parts of PUS movement of goods and people is westward to Mamuju; farther south it is toward Polewali either rather more directly or through Malabo. The road from Malabo to Mambi was once open to vehicular traffic, perhaps only four-wheel drive. It is now passable to some nine kilometers from Malabo. Continuing on toward Mambi, there is usually a two-track trail, but at various spots, perhaps totalling 10% of the way, it is currently impassable to cars. There was evidence of recent bulldozer work on the road. We were told that there is a more direct road from the south coast to Mambi currently under construction. Coming in from the west, there-is a road from Kaluku on the coast to Kean. Merpati flies out of Kaluku twice a week. This area is under development, we were told, in order to service interior Kalumpang. If that is the case, it seems unlikely that there will be any development on the PUS side of Kean soon. Movement within the PUS area 1S either by foot or by pack horse. Bousing, research possibilities In a number of cases it was clearly stated to us that a longer tera stay by resident linguists would be welcome. In some villages such a stay would require building a house for a family 14 PUS1 com~ng in from outside the PUS area. In other villages we were told a house could be rented. Allocation factors The single largest allocation factor 1S the question of dialects. We are happy, on the one hand, that the seven dialects of the Grimeses' report have been reduced to potentially three, howbeit subject to much more testing: Aralle-Tabulahan, Bambang, Rantebulahan-Maha1a'an. On the other hand, the greater linguistic distance between (at least two of) the three, make allocation decisions a bit stickier. Tabu1ahan seems to be regarded in some sense as the original or central dialect (mythological origins derive from Tabu1ahan). Bambang, however, is more basic, on first glance, to the political, social and economic directions of the PUS area. Another fac~or in deciding on placement is the distribution af Muslim, Christian and animistic villages. The Bambang dialect is spoken by some of all three, and probably would give opportunity potentially to interact with all three subgroups. The extent that Aral1e-Tabu1ahan and Rantebu1ahan-Maha1a'an is understood by all three religious groups still needs to be determined. Also to be considered is the presence of Ken -George in Mambi through June, 1985. He requested that SIL not place a team in the area of his jurisdiction, that is, in desas Mambi and Bambang. Finally, isolation cannot be overlooked. Without an airstrip at least for emergency use, any UNHAS-SIL team would be at least one day's trail time from a main road, in most places two days. Lakahang, the least isolated of northernwPUS locations and flat enough to have its own airstrip, has the double disadvantage of generally being on the fringe of the PUS area and of being too cosmopolitan, too heterogeneous. Of the PUS villages which we either passed through or where we actually stayed a night (Mambi, Aral1e, Saludadeko, Tampalopo, Tabu1ahan, Taora, Sodangan), any of them but Mambi would be good locations for an UNHAS-SIL team, at least for the first year or until the area had been evaluated in greater depth. PUS1 15 B.eferenees GrDDes, Charles and Barbara Gr~es, Languages of South Sulawesi (to appear). Mandadung, Arianus, 1982. Mamas a (Kondosapata' Waisapalelean) Dalam In£ormasi-Sejarah, -Budaya, -Pariwisata. Samar, Abd. Azis and Arianus Mandadung, 1979. Ungkapan Sejarah dan Budaya di kabupaten Po lewali-Mamas a , SulSel. Seri "A" - (Daerah Kondosapata'!Mamasa). 16 PUSl 1JRIIAS-SIL SOCIOLIIICIJISTXC SuKVSI: XABUPATER POLEWALI MAMASA, WEST-c:aDAL SBCnOR Kare J. Str~mme TABLE OF CONTENTS Page INTRODUCTION 18 THE TERM "PlTU ULUNNA SALU" 18 SURVEY ITINERARY 19 LANGUAGE ANALYSIS AND RESULTS 24 OUR FINDINGS COMPARED WITH THOSE OF THE FORMER SURVEYS 32 PRESENT RESIDUES - FORTIlERRESEARCH 3S LANGUAGE USAGE 3S NONLINGUISTIC INFORMATION 36 CONCLUSION 39 REFERENCES 40 Table 1: Government and Population of Area Surveyed 22 Table 2: Cognate Percentages and Grouping of Wordlists 27 Table 3: Languages and Dialects Surveyed with Estimated Number of Speakers 29 Figure 1: PUS Dialect Chain 32 Map 1: Kabupaten Polewali Mamas a , West-Central and Northern Sections 21 Map 2: Languages and Dialects of Kabupaten Polewali Mamasa, West-Central and Northern Sections 28 PUS2 17 DITBODOCTIOB The survey was conducted by Thomas V. Laskowske and Kare Str~e September 27 October 9, 1984 with the intention of complementing earlier UNHAS-SIL surveys in the area, i.e. the initial survey carried out by Charles E. Grimes and Francis B. Dawson in January 1983 and the following survey by Timothy Friberg and Eui Jung Kim in December 1983, thus bringing the general survey of the whole area to completion. The survey conducted by Grimes and Dawson gave an initial general overview of the language situation in this area, often refer~ed to as pitu Ulunna Sa1u (PUS), based on word1ists taken in the village of Mambi only. Friberg and Kim carried out a more detailed survey of the northern part of PUS, seeking to verify the conclusions made by Grimes and Grimes in Languages of South Sulawesi (to appear). In his pitu U1unna Salu survey report of kabupaten Polewali Mamas a , Northern Section; Friberg basically agrees with the conclusions of Grimes and Grimes for this northern part, although he seriously questions the Grimeses in taking Aralle-Tabulahan and Bambang as dialects of one language instead of two separate languages. Friberg also expresses several opinions about the language situation in the central and southern parts of the PUS area, but these are only based on second-hand information. Our immediate purpose for this survey, therefore, was to investigate the linguistic situation in central and southern PUS, expanding on the work of Grimes and Dawson and, together with the findings of Friberg and Kim, to establish more exactly the dialect boundaries in the PUS area. In doing so we visited kelurahan Mambi and desas Rantebu1ahan and Mehalaan in kecamatan Mambi as well as neighboring desa Sindagamanik in kecamatan Mamasa and Sumarorong village to the east and desas Tapango and Bulo in kecamatan Wonomulyo to the south. (Kecamatan (subdistrict), the administrative unit between kabupaten (district) and desa (a smaller administrative area made up of several villages). The term kelurahan, which corresponds to desa, is used for more central/prestigious areas.) Through processing the 14 wordlists and 10 sociolinguistic questionnaires that were obtained, all except Rantepa1ado, Matangnga, and Bu10 in the respective villages, the objectives for the survey were met. TIlE TBJD( "PlTU lJLDllIRA SAW" The term IIpituUlunna Salu" literally means "seven river head". According to historical information available about the area, "Pitu" (seven) refers to seven members of a former sociopolitical federation, while "U1unna Salu" (river head) designates the area of this federation, i.e. the area around the upper part of the Mapili river. The federation, which was formed 18 PUS2 in the distant past and stayed more or less intact until the Dutch took over the administration of the area, consisted of the following seven areas: Tabulahan, Ara11e, Mambi, Bambang, Rantebulahan, Matangnga, and Tabang. Except for Tabang (northeast of Mamasa) all of these areas are within the area covered by the different PUS surveys, and except for Tabang and Matangnga all of them are within kecamatan Mambi. The people in this area, however, never seem to have referred to themselves as the PUS people but always as "To" + name of village (e.g. "To Aralle"), which is still the case. Moreover, nobody seems to refer to his language as the PUS language. To name their language they either use "Bahasa" + name of village or a local term. This was probably also the case in former times. According to our findings the parts within this area that are most remote from each other (disregarding Tabang, for which we have no information), relate to each other linguistically at a cognate level of only about 70%. This relatively high difference probably explains why the people prefer to name their language according to their village rather than using a common language name for the whole area. Nevertheless, many linguists, including Esser, Salzner, and Grimes and Grimes, have found the linguistic unity of the area sufficient to speak of one language, for which they have used the name of the former sociopolitical federation, Pitu Ulunna Salu. (It should be noted that Grimes and Grimes also have included within the PUS language two dialects, Tapango and U1unda, that are not spoken within the limits of the former PUS federation. Interestingly, we have found that these dialects do not belong to the languages spoken within the area of this federation.) As Pitu Ulunna Salu is already a commonly accepted label in linguistic circles, and in the absence of better alternatives, we will continue to use this label for the largest language of the area. Because we have also suggested another language within thiR area, Ara1le-Tabulahan, it should be noted, however, that our use of PUS as a language name is not identical with that found in previous descriptions of the area. Our use of PUS as a label for a subfamily of languages is also new. The area covered by this subfamily is somewhat larger than the area of the former PUS federation. SDKYKt ITIlIER.AB.Y Sept. 27: We left Ujung Pandang on motorcycles in the afternoon and drove to Pare-Pare, where we spent the night. Sept. 28: We drove to Po1ewali early in the morning. Reporting at the bupati's office (office of the highest administrative officer at the kabupaten level), we met with the PUS2 19 Sekwilda (Sekretaris Wilayah Daerah, the second highest administrative officer at the kabupaten level), who immediately wrote letters of introduction for us to the camats (highest administrative officer at the kecamatan level) of Mambi, Sumarorong, and Wonomulyo. As the camat of Mambi was in Polewali at that time, we went to meet with him after lunch. He wanted to write the letters of introduction to the kepa1a desas (highest administrative officer at the desa level) in Mambi himself, but he had to contact his office in Mambi via radio before he was able to do so. We spent the rest of the day in Polewa1i. Sept.~: The letters were ready about 10 a.m., and we left immediately for Mambi, stopping briefly in Sumarorong, where we left a letter for the camat. We arrived at Malabo at 4.15 p.m. after an additional stop because of heavy rain, and then started out on the very poor and slippery road from there to Mambi. After a hard drive with several falls, we reached the village of Tondok Bakaru just before dark. The kepa1a kampung (administrative head at the village level) there kindly invited us to stay in his house for the night, which we gladly accepted. We also took a word list from him that evening. We asked about the condition of the road from there to Mambi, but the answers were somewhat unclear and conflicting. Sept. ~ The trail was still quite muddy next morning, so we first decided to leave our motorcycles there and walk in to Mambi. We hired a horse to carry our baggage and set off. After a short distance we met a man just coming from Mambi. He assured us that the trail was in good condition all the'way, so we were convinced to go back to get our motorcycles. But we did not get far on our bikes before we realized otherwise. After 9 km and 2 hours of hard struggle, we decided to park the cycles again in the village of Galung-Ga1ung. It took us about 3 hours to walk the 12-13 km from there to Mambi. In Mambi we stayed at the house of Pak Rahman, a man we had met in Polewali who had invited us to stay at his house. We were able to use this house as a base for several of our trips the following days. When we arrived, we met with Ken George, an American anthropologist who was staying in the same house. He arranged for us to meet with the local police that night and gave us helpful information about the linguistic situation of the area. We were also able to get a wordlist from a native Mamb1 speaker that evening. Oct. 1: The police came to check our papers again, and in the afternoon we got the permission to conduct the survey. Meanwhile Ken George had asked us not to enter desa Bambang for various reasons relating to his research work. When he was able to provide a native Bambang informant (from the Rantepalado village), we agreed to this and took our Bambang wordlist with h~. 20 PUS2 Map1: KabupatenPolewali Kamas. West-Central and Northern Section LEGEND. Kab. border Kec. • Des. • Vehicular road Foot/Horse trail A~inlstrative town of Kab. • village of Kec. • •• Desa Other II i 11ago River Distanc. b.tw •• n \/111ag•• in k~. N••• Dr ad.\nl,tr.ti". unit Na•• Dr "i1lag. and ad.inia- tratlv. urrI t PUS2 21 Table 1: Government and Population of Area SUrYeyed Kecamatan Pop. Desa/Kelurahan Pop. Village surveyed Pop. Mambi 31846 Buntumalangka 5416 Tampalopo 225 (Burnal) Saludadeko 311 Tabulahan 2404 Tabulahan 400 Periangan 225 Aralle 5341 Aralle 200 Bambang 5828 Rantepalado 600 Mambi 5586 Mambi 200* Galung 100 Rantebulahan 3890 Keppe 445 Ga lung-Ga lung 200 Mehalaan 3381 Mehalaan 1000 Kondo 100 .Mamasa 30145 Sindagamanik 2124 Tondok Bakaru 860 Sumarorong 16061 Tadisi 2626 Sumarorong 500 Matangnga 2599 Matangnga 1000 Wonomulyo 76568 Tapango 4088 Tapango 535 Bulo 5125 Polliwa 500 Bulo Not reported Karomban 90* Notes: Population figures for the kecamatans and desas/kelurahans are taken from Kabupaten Polewali Mamasa Dalam Angka~. (Kantor Statistik kabupaten Polmas). Population figures for the villages are estimates given by our informants. * Indicates reported number of households. 22 PUS2 Oct. 2: We walked to Mehalaan, 14 km - 3 1/2 hours' walk, and did a wordlist and a sociolinguistic questionnaire there. We were informed there that the dialect at Kondo, the southernmost village of the desa, was quite different from that at Mehalaan. Unfortunately, we had already arranged to go back to Mambi that night, stopping for a wordlist and questionnaire at Keppe on our way, so it was not possible to continue to Kondo then. We went on to Keppe, completed our work there with the kepala desa, and got back to Mambi at about 8 p.m. Oct. l: We left Mambi for Galung, the westernmost village of the kelurahan, with a guide in the morning. After 8 hours including numerous stops and changes of guides in the villages along the route, we finally reached our destination (actually only 16 km from Mambi). As the kepala kampung was not present, we were received by the religious leader in the village, and we were able to complete a wordlist that night. Oct. 4: After we had also completed a questionnaire at Galun~ w; returned to Mambi. This time we were able to do the distance in about 5 hours. Having hired horses to take our baggage, we immediately set off for Galung-Galung in the eastern part of Rantebulahan, where we had parked our motorcycles. We completed a wordlist with the kepala kampung there the same night and decided to go on to Kondo the next day. Oct. 5: Having completed a questionnaire in Galung-Galung, we ser- off for Kondo, this time on our motorbikes. We had been told that the trail from there to Mehalaan was quite good, which proved to be true. We were able to drive the 15 km in about an hour. We parked our motorcycles in Mehalaan, got our letter of introduction from the kepala desa's office, walked the 18 km to Kondo, and got a wordlist and a questionnaire there the same night. Q£!.~: We walked back to our motorcycles in Mehalaan and drove the S4 km from there to Sumarorong. The stretch from Mehalaan to Sika via Leko was reasonably good for motorcycles, much better than the Mambi trail. Arriving at Sumarorong at about 8 p.m., we checked in with the head of the camat's office and spent the night at a guesthouse nearby. Oct. 7: The head of the camat's·office arranged for us to meet -;Ith- a native Sumarorong speaker in the afternoon, and we were able to get both a wordlist and a questionnaire filled 1n. We were also able to make contact with some people from Matangnga that evening, and arranged to come back next morning to do a wordlist and a questionnaire with them. Oct. 8: Having finished our work with the Matangnga people, we dro;e down to Polewali, where we checked in with the kabupaten police. We then went straight on to Wonomulyo, checked in with the kecamatan police and the camat there and set off for Tapango (on motorcycles) with a guide from the camat's office immediately PUS2 23 after lunch. We got both a wordlist and a questionnaire filled in with the help of the kepala desa there and went on to Polliwa in Desa Bulo with the same guide. There we met with the kepala desa, who also served as our informant. He had grown up in the nearby village of Bulo, but claimed to be aware of all the dialect differences between the two villages, so we actually made two wordlists based on his information. The night was spent in his house. Oct. 9: Being somewhat uncertain about the accuracy of the Polliwa-wordlist taken the night before, we checked it with a native speaker from that village. Then we drove on to Karombang, the most interior village of desa Bulo accessible by motorcycle. Having completed a wordlist and a questionnaire there, we headed back to Wonomulyo, checked out at the camat's office and the police there, and drove back to Ujung Pandang that afternoon and evenlng. LAlIGDAGE ABALYSIS on llESULTS Our survey objectives have basically been attained by a comparison of the 14 wordlists we took in the area. To get the overall picture of the language situation in the whole PUS area, we have also included in our comparison the lists taken by Grimes and Dawson and by Friberg and Kim on their survey trips to this area. We have also included several lists taken by Grimes and Dawson in the surrounding areas, i.e. the Mamasa, Pattae', and the Balanipa and Malunda lists of the Mandar language. The 10 sociolinguistic questionaaires taken and informal questioning of various people, have shed further light on the language situation. The wordlist used was the same that was used by Friberg and Kim and almost identical to the one used by Grimes and Dawson on their PUS surveys. Of the 216 Indonesian words on the list, 16 were eliminated for the comparison, many because the words caused confusion and were not answered satisfactorily. The cognate decisions were made according to the inspection method (synchronic approach). This means that the comparison was made on the basis of present phonetic similarity, taking into consideration, of course, features such as consistent correspondances, as for instance between [h] and [w] and between [h] and [r] in various dialects. The same method was used both by Grimes and Grimes and by Friberg in their cognate decisions. We also made our own cognate decisions on the lists previously taken by Grimes and Dawson and by Friberg and Kim in order to make sure that the same methods were applied to all lists included in our comparison. Our cognate percentage calculations turn out very similar to theirs, with an average deviation of only +/- 1.4% from Grimes and Grimes' and + 2.7% from Friberg's 24 PUS2 calculations. On that basis we may conclude that our methods in making the cognate decisions have been quite similar, although it is evident that we have been somewhat more liberal than Friberg and Kim in this respect. Grimes and Grimes have grouped their word lists according to the following hierarcy on the basis of shared cognate percentages: Under 15%: Belong to different phyla Over 15%: " " the same phylum 11 25%: " " It " superstock " 45%: " " " " stock " 60%: " " 11 " family II 75%: " n 11 11 subfamily " 80%: " " " " language These levels, however, cannot be followed blindly when grouping wordlists. Other factors must also be taken into consideration, such as how the people in the area view the relatedness of their language with those around them. To make more definite decisions about language boundaries may require some intelligibility testing. In our grouping we have basically followed Grimes a~d Grimes' 80% limit in dividing up the languages. The Tondok Bakaru list, however, we have chosen to group with Mamasa due to the fact that this is clearly how the people there perceive the relatedness of their dialect, although it also relates to the PUS lists at 80% or above. Also the cognate percentage is somewhat higher with Mamasa than with the PUS lists. Mambi relates both to the Aralle-Tabulahan and the PUS lists above 80%, but as it shows a somewhat higher lexical similarity with" the former group (86% with Grimes and Grimes' Ara11e-Tabu1ahan list), we have chosen to include Mambi as a dialect of the Aral1e-Tabulahan language. It should not be overlooked, however, that Mambi has a clear transitional position between the Ara1le-Tabulahan and PUS languages. The same can be said about Matangnga in relation to PUS and Pannei. Despite its clear transitional position between the two, we have chosen to include Matangnga as a dialect of PUS because of its somewhat higher average lexical similarity with the PUS lists. In establishing the PUS subfamily we have chosen to go down from Grimes and Grimes' 75% limit to about 70%. There is a clear chaining relationship between the languages grouped into this family, and there is also a clear drop in cognate percentages between this group and the Pattae' and Mandar languages. As for the lists grouped under the Mamas a language, these share a cognate average of about 79% with the PUS lists. It is therefore well understandable that Salzner could group these together in a Sa'dan group. Despite this we have chosen to follow PUS2 25 Esser and Grimes and Grimes in separating the two. When the Aralle-Tabulahan and Pannei languages are included in a PUS subfamily, Mamasa definitely relates closer to the other languages Gr~es and Gr~es list under the Toraja-Sa'dan subfamily than to the PUS subfamily. Gr~es and Gr~es do not clarify what cognate percentage they have used to distinguish between dialects of one language. All their word lists taken in the PUS area are listed as separate dialects. As the highest cognate percentage between any of these lists is 91%, we may conclude that they have considered anything under that figure to be separate dialects. We have basically used that same figure in separating the dialects. Due to the extensive chaining pattern in this area, however, it is very difficult to draw boundaries. Extensive intelligibility testing will have to be undertaken in the area before firmer conclusions may be drawn. As our survey did not include any testing of that kind, our decisions with regard to the dialect boundaries must be considered to be somewhat tentative. Based on the criteria described above, the results of our analysis are portrayed in the matrix of Table 2. These may be summarized as follows: Separated from the Toraja-Sa'dan subfamily to the north and east and from the Pattae' and Mandar languages to the south and west, we suggest a PUS subfamily of languages. This consists of the PUS language, the Aralle-Tabulahan language, the Pannei language, and probably at least one other language to the west of these, represented by our wordlist from the village of Kondo. As for the dialect divisions within these languages, we tentatively suggest eight dialects within the PUS language (i.e. Bumal, Issilita', Bambang Hulu, Salu Mukanan, Mehalaan, Pakkau, Pattae', and Matangnga), three dialects within the Aralle-Tabulahan language (i.e. Aralle, Tabulahan, and Mambi), and two dialects within the Pannei language (i.e. Tapango and Bulo). In the following paragraphs we will further explain these conclusions. Mambi and Matangnga's special pos1t10ns have already been commented on above. Due to the extensive linguistic chaining in the area, it is also difficult to group the other lists. The picture is further complicated by the fact that we do not know exactly what villages Grimes and Dawson's informants represented, as their wordlists are only named after the desas (Bambang, Rantebulahan, and Mehalaan). Also we were not able to check out the villages in desa Bambang in person. (Our Rantepalado wordlist (western Bambang) was obtained from an SMA student in Mambi). According to Ken George and information obtained at Galung-Ga1ung there are two dialects in desa Bambang, Issilita' (the western part, including Rantepalado) (Losudabota, as our Rantepalado informant called his dialect, is probably an alternate name for 26 PUS2 Table 2: Cognate Percentages and Grouping of Vordli8ts PUS2 27 Kap2: LaD.guapa an4 dialects of kabapateaPol_li __ a West-Central and Northern Section Language boun~ary Oialect boundary I:ao. border Village where wordlist has been taken Other town l.anguage namo Name of dialect Name of village and dialect 28 PUS2 Table 3: t.anaaaPa aacl Dialect. of Area SU1wet'ed with .ati..at:ed ~ of Speakers Language Dia1ec t PUS 22000 Buma1 5500 Issi1ita' } 6000 Bambang Hu1u Sa1u MUkanan} 4000 Pakkau Meha1aan 3500 Pattae' 500 Matangnga 2500 Ara11e-Tabu1ahan 12000 Ara11e 5500 Tabu1ahan 2500 Mambi 4000 Mamasa 50000 Sindagamanik 2000 Sumarorong 10-15000 ? "Kondo" 500 Pannei 9000 Bu10 5000 Tapango 4000 -Note: The figures above are estimates made on the basis of the population figures given for the roughly corresponding administrative units in table 1.o PUS2 29 Issi1ita') and Bambang Hu1u (or To Sa1u)(the eastern part). As our Rantepalado word1ist relates to the GrLnes and Grimes' Bambang list at only 90%, it is quite likely that these represent those two different dialects. There are also cases of consistent phonetic differences between the two lists. Thus the Bambang list has [h] where the Rantepalado list has [r]. We have therefore decided to follow what was reported about the language situation in Bambang, splitting the desa up in two separate dialects. The question still remains, however, how these dialects relate to those to the north and to the south. According to information obtained in Mambi and to some extent in Galung-Galung, due to a former migration of people fr~m the south to the north in this area, the villages in eastern Bambang share the same dialect with those in the eastern part of Rantebulahan to the south and with those in desa Bumal to the north. This is not evident from our cognate percentage calculations, however. According to these both the Rantepalado and the Bambang lists relate to those from Bumal (Saludadeko and Tampa1opo) at only 90%, and to the Ga lung-Galung list (eastern Rantebulahan) at 91%. This seems to indicate that the two Bambang dialects may be distinguished from both Bumal and eastern Rantebulahan. As Saludadeko and Tampalopo relate to each other at 94%, it would seem natural to group them into one Bumal dialect distinct from the Bambang dialects. But as to eastern Rantebulahan the picture is more obscure. Galung-Galung relates to our Meha1aan list as high as 94% (92% with Grimes and Grimes' Mehalaan list, which probably represents a village closer to Keppe). The two Mehalaan lists relate to each other at only 90%. This would seem to indicate that eastern Rantebulahan should be included with Mehalaan rather than with eastern Bambang. However, if we take some of the phonetic features into consideration, these seem to point to the opposite conclusion. Thus the Galung-Galung list groups with eastern Bambang in having [h] where Mehalaan has [r]. Also in Galung-Galung all final nasals are [m] as on the Bambang list, while in Meha1aan they are [n]: or [n] , So this case obviously needs to be further checked. First of all the word1ists already taken ought to be double checked, adding missing synonYms, for instance. More word1ists from the eastern part of Bambang (assuming Grimes and Grimes' list is from that part) would probably also shed further light on the situation. Meanwhile, because of the incomplete and ambiguous data available and also because of the fact that people at Ga1ung-Ga1ung gave a separate name for their form of speech, Sa1u Mukanan, we have chosen to call the form of ,speech in eastern Rantebulahan a separate dialect. We have chosen to name this dialect Sa1u Mukanan, in accordance with our Galung-Ga1ung informant. Keppe relates to Grimes and Grimes' Rantebulahan list at 93%, 4% higher than to Galung-Galung and 3% higher than to Grimes and 30 PUS2 Gr~esf Mehalaan list. This,seems to indicate that the informant for Grimes and Grimes' Rantebulahan list was from western Rantebulahan and that this area forms a separate dialect. The dialect split through desa Rantebulahan was also reported at Keppe. The division corresponds to the division between Christianity and Islam in this desa, going through the village of Keppe. (Our wordlist was taken in the western, Muslim part of the village.) In accordance with information given at Keppe, we have chosen to call the dialect of western Rantebulahan Pakkau. Galung, in the western part of kelurahan Mambi, relates closest to the Grimes and Grimes' Rantebulahan list and our Keppe list, 91% and 89% respectively. This further indicates that the Grimes and Grimes' Rantebu1ahan list probably belongs to western Rantebulahan, most likely to the area west of Keppe. Despite the clear ties with western Rantebulahan, we have chosen to suggest a separate dialect for the area of Ga1ung and to call it Pattae' (not to be confused with the Pattae' language, also included in our matrix) in accordance with the local people. According to information obtained there, the dialect also includes the villages of Salung and 5ambaho in the western part of kelurahan Mambi, as well as the villages of Urekang and Pupenga across the border to kabupaten Majene (in desa Ulumandak). As illustrated in figure 1 below, there is a clear chaining relationship between these dialects. All relate to the neighboring dialects at 89-90%, while the extreme points in the chain only relate at 84-85%. Although included as a dialect of the PUS language, Matangnga has not been included in the dialect chain of figure 1 as it -does not seem to be in a clear chaining relationship with these other dialects. But, as has already been pointed out, Matangnga forms an important link between the PUS and Pannei languages. Our matrix of cognate percentages also seems to point to some kind of chaining relationship between Aral1e-Tabu1ahan, Kondo, and Karombang. Thus the Kondo wordlist, as the central link, relates to both the Karombang, ,Grimes and Grimes' Aral1e-Tabulahan, and Friberg and Kim's Aralle wordlists at an average of 79%, while these Aralle-Tabu1ahan lists relate to Karombang at an average of only 15%. The close relationship between these western villages of the area, together with the clear gap between these and the Mandar dialects, indicates some kind of linguistic unity in this intermediate area. This also corresponds with the information obtained at Kondo and Galung. At Kondo the residents told us that they were Umnigrants from Lakese in desa Bulo and that they speak the same dialect as the people in that village and Tanete (also in Bulo). Penatangan (desa Bulo) and Tubbi (desa Tubbi, kecamatan Tuta1lu) were said to share a closely related dialect. Interestingly, people at Kondo PUS2 31 Bumal ~ Issilita' Bambang Hu It -.. Salu Mukanan Pattae' Pakkau Mehalaan Pigare 1: PUSDialect Cltain also called their language Pannei. This might be a further indication of chaining in the area between Kondo and Karombang. At Galung we were told that Bahasa Ulumandak is the same as Pannei. This indicates that the same chain continues northwest from Kondo into desa Ulumandak in Majene. Although different from their dialect, people at Kondo claimed to understand Ulumandak. According to information Friberg and Kim obtained at Aralle, the people there also claim Ulumandak to be a closely related dialect. This points to the continuation of the chain into the area of Aralle. To check all these claims and indications, investigation needs to be undertaken in the desas of Ulumandak and Tubbi as well as the northern part of desa Bulo. Meanwhile we have chosen to separate Kondo from both the Pannei and the Aralle-Tabulahan languages. OUR FIllDIBGS COIIPAllED trIm mon OF THE I'OJlMD. SURVEYS In the following we will briefly compare our conclusions with those made by Grimes and Grimes and Friberg and Kim for this area, pointing out' similarities and differences. Later we will sUmmarize the new findings that have come of of our survey. 32 PUS2 Si.ilarities We have confirmed Grimes and Grimest classification of PUS as distinct from both Mandar and Toraja Satdan. We have also confirmed their observation that the PUS dialects relate to each other in a complex pattern of interrelationships, mainly by the feature of chaining. As far as the actual dialects are concerned, Meha1aan is the only one that we both list with the same name as a dialect of PUS. The other dialects they list, we have either grouped outside of the PUS language or divided up as two dialects with different names. We have confirmed the existence of a Tapango dialect, but have not included it within the PUS language, rather as a dialect of Pannei. As far as Ulunda is concerned, we think that what the Grimes and Grimes really refer to is the U1umandak dialect spoken across the border to Majene. According to present information this dialect may not be classified as a dialect of PUS. It therefore appears that the PUS dialect we found in the western part of kelurahan Mambi, Pattaet, has until now been overlooked. With Friberg and Kim we agree with separating out Periangan from the PUS language. Concerning Aralle-Tabulahan they say that they are in serious doubt as to whether this area can be considered a dialect of PUS due to the low cognate percentages with the other PUS dialects (except Mambi). We have acted on this doubt in suggesting a separate Aralle-Tabulahan language. Regarding the internal relationship between Aralle and Tabulahan, we agree with Friberg and Kim in describing them as close but separate dialects. Whether they may be reduced to one dialect, as they have indicated, still remains to be proved. The same is the case with regard to the Bumal and Bambang dialects and the Rantebulahan and Mehalaan dialects, of which both sets show an internal cognate relationship of about the same level (i.e. 89-90%). Although Friberg and Kim did not take any wordlist ~n Mambi itself, they observed that Mambi is not closely related to any of the other dialects of the PUS area. Ours and Grimes and Grimes' wordlists from Mambi confirm that. The closest relationship is with the Aralle dialect at 86%. Differences As already indicated, our conclusions differ to some extent with those of the former surveys as to which dialects should be included within the PUS language and as to how the dialect divisions within PUS should be made. Contrary to Grimes and Grimes we have separated out Aralle, Tabu1ahan, Mambi, and Tapango, classifying the former three as dialects of a separate Aralle-Tabulahan language, and the latter as a dialect of the Pannei language. As already noted, we have found a new dialect, Pattae', in the western part of kelurahan PUS2 33 Mambi and explained Ulunda as referring to Ulumandak ~n Majene. Their Bambang dialect we have divided into three dialects, Bumal, Issilita', and Bambang Hulu. Rantebulahan we have divided into two dialects, Pakkau and Salu Mukanan. As they did not obtain any wordlists from Bambang, Rantebulahan, and Mehalaan, Friberg and Kim only suggest a possible configuration for these areas. They talk about a Bumal-Bambang related set of dialects and a Rantebulahan-Mehalaan related. set, without mentioning anything about dialect chaining. Our findings do not support this grouping. According to our data the border between Bambang and Rantebulahan (eastern part) is no more clear than the borders between the other adjacent links in the dialect chain we have suggested. As Figure 1 shows, there is a continuous chaining all the way from Bumal to Pattae', each dialect relating to the adjacent dialects at about 90%. Thus, rather than to related sets of dialects, the current data point to a series of separate dialects in a Chaining relationship to each other. Due to lack of data Friberg and Kim also were not aware of the dialect division within desa Bambang and the division between the eastern and western part of Rantebulahan. Although it may be possible to combine some of the dialects after intelligibility testing, our eight PUS dialec~s are not likely to be reduced to the three postulated by Friberg. Hew Findings The new findings resulting from our survey can be summarized as follows: - Two dialects, Issilita' and Bambang Hulu, within desa Bambang. Two dialects, Pakkau and Salu Mukanan within desa Rantebulahan. A separate PUS dialect, Matangnga, in the western part of kecamatan Sumarorong. A separate PUS dialect, Pattae', in the western part of kelurahan Mambi (assuming that Ulumandak is what Grimes and Grimes "mean by Ulunda)." A separate language to the south of PUS, Pannei, of which Tapango is a dialect. A speech form, represented by at least the village of Kondo, sufficiently .different from all of the other languages in the area not to be included with any of them, thus indicating at least one other language in the area southwest of PUS. 34 PUS2 PRESENT RESIDUES - FOlUiWl URBCIl We have already referred to several difficulties still needing investigation in order to arrive at firmer conclusions regarding some aspects. These may be summed up as follows: We are somewhat uncertain as to whether eastern Rantebulahan (Salu Mukanan) should be classified as a separate dialect or grouped with Mehalaan or with Bambang Hulu. To shed further light on this more word lists should be obtained from eastern Bambang (taken within the area), and the Galung-Galung list should be further checked for synonyms. - We are unsure where the language boundaries of the language represented by Kondo are, we do not know the precise extent of the Pannei language, and we are generally unsure about the linguistic situation in the area between PUS and Mandar (desa Tubbi and the northern part of desa Rulo in Polewali Mamasa and desa Ulumandak in Majene). A linguistic investigation of this area is needed to answer these questions. We are somewhat unsure of the necessity for establishing eight different PUS dialects. In order to check this, extensive intelligibility testing needs to be undertaken, especially between the dialects included in the chain of Figure 1. The same is true with regard to the dialects of the Ara1le-Tabulahan and Pannei languages. - We are not sure how the Tabang area, a member of the former PUS federation, relates linguistically. A wordlist sho~ld be taken there and be compared with both the PUS, Aralle-Tabulahan, and Mamasa languages. LARGUAGE USAGE As we did not do intelligibility or bilingualism testing or any extensive observation of language use, the following comments are based on what was reported to us as we filled ~n our questionnaires and on casual observation. Throughout the area visited, Indonesian is claimed to be known and used in addition to the local language. Only in a couple of places our informants told us that their knowledge of Indonesian was limited, and in one village (Karombang), although not told so, this was obviously the case. How widespread the knowledge of Indonesian is within the various villages needs to be further tested, especially as our informants were all adult males, most of them having some official position and/or a reasonable education. At several places the parents of our informants were reported not to know Indonesian. Obviously the introduction of schools, even in the most remote areas, has changed the situation a lot. PUS2 35 At several places languages of larger neighboring groups such as Mandar, Bugis, and Mamasa were claimed to be known. In most eases it was stated that this was an acquired language ability due to frequent contact with these areas, and that the languages were not mutually intelligible. With regard to claims of usage it was universally claimed that the local language is the one normally used within the family. Also when working in the rice fields, rebuking other people, holding traditional feasts, or telling traditional stories, the local language is the one used. At places such as government offices and health and religious centers both Indonesian and the local language are used in most cases. In some of the less isolated areas, however, only Indonesian is reported to be used at these places. We may safely conclude that the local languages play an important role in the daily lives of the people in this area. Further investigation needs to be done to establish the extent of the Indonesian knowledge in its various communities. RDlILIRGUlsnc IRPOKMATIOR Although with regard to the linguistic situation we have sought to give an overview of the whole area by also including data from the previous surveys, apart from the historical section, we will concentrate on the particular areas we visited for the following information. A lot of similar information for the northern part of PUS can be found in Friberg's survey report. History Some historical information is available in the books Mamas a , Da1am Informasi, -Sejarah, -Budaya, -Parawisata by Arianus Mandadung (1982) and Ungkapan Sejarah dan Budaya di Kabupaten lt Polewali-Mamasa,' Su1-Se1 Seri:· "A 2 (Daerah Kondosapata'/- Mamasa) by Abd. Azis Samar and .Arianus Mandadung (1979). The following is a short summary of this information. According to tradition the forefather of the people in the area, Pongka Padang, moved from Toraja and settled in the Tabulahan area some time in the distant past. He and his seven descendants later spread to all the places that eventually became members of the PUS federation. In this early period there were relatively frequent conflicts and fighting between the various villages as well as with the surrounding peoples. Probably due to the evident destructive effects of this fighting and increasing pressure from outside the area, the formerly mentioned sociopolitical federation was established between Tabulahan, AraIle, Mambi, Bambang, Rantebulahan, Matangnga, and Tabang. Each member in the federation was assigned specific tasks or roles. This relatively united federation proved quite effective, 36 PUS2 remaining more or less intact until the Dutch took over the administration of the area. It was not formally broken up at that time either, but naturally lost most of its power and influence. Under the Dutch the PUS-Mamasa area was made a subdivision of the Mandar administrative area, with the subdivision government situated at Mamasa. This area was further divided up into three subdistricts, of which Tabulahan, Aralle, Mambi, Bambang, Rantebulahan, and Matangnga formed one, administered from Mambi. After the Dutch rule Mamasa (including PUS) continued to be part of a Mandar administrative area until 1959. Then this area was divided up into three kabupatens, i.e. Polewali Mamas a (Po1mas), Majene, and Mamuju. The former Mamasa subdistrict was divided into four kecamatans, i.e. Mamas a , Pana (northeast of Mamasa), Sumarorong, and Mambi. Geography As with the north visited by Friberg and Kim, the southern part of the PUS area is basically rugged and mountainous. The only level areas of any size are the rice fields around Mambi, Keppe, and Meha1aan. Towards the western and southern part of kecamatan Mambi the elevation declines, but the landscape is still very rugged, so rugged that rice paddies are a rare sight. On the road into the PUS area from Malabo, the highest point to be passed is on the border between Mambi and Mamasa kecamatans. From that point there is a continuous descent both to the west and the southwest. All the rivers to the west of this point are branches of the Mapili river, draining towards the south. Apart from the cultivated areas around the villages and a few outlying dry-field areas, most of the area is covered with forest. The area we visited in kecamatan Wonomulyo to the south has quite a different geography. This area is just off the coastal plain, so there are no real mountains. In Tapango there are only relatively small hills; in desa Bulo it gets somewhat more rugged and the forest gets thicker. Econoay Subsistence farming is the livelihood of most people in the southern region of PUS, rice being the most important crop. Wet field cultivation dominates where there are sufficiently level areas. Due to the lack of such areas towards the western and southernmost parts of kecamatan Mambi, dry field cultivation is common there. Cassava is also a common crop in those areas. Otherwise various kinds of vegetables are commonly grown throughout the region. Coffee seems to be the most common cash crop, usually being sold to traders at the Mambi or Galung-Galung markets by those in kecamatan Mambi. Some rattan and other wood products are sold. Basic food supplements, clothing, and var~ous tools are bought with the cash. Meat seems to be rare in the PUS2 37 diet; dried fish ~s more common, usually bought at the local markets. Education Most of the villages in the area (even isolated ones like Botteng and Kondo) have primary schools (SD). Mambi and reportedly Leko also have SMPs (secondary schools), and an SMA (high school) has recently been opened in Mambi. In the coastal area there are SMAs in Po1ewa1i and Wonomulyo. These are also attended by many students from kecamatan Hambi, especially from its southern parts. The educational facilities in the area must be considered relatively good. Religion The area is divided between Christianity and Islam. Christianity is the dominant religion in desa Bambang and the eastern part of desa Rantebulahan. The boundary between Christianity and Islam in Rantebulahan goes through the village of Keppe, the eastern part being Christian and the western part Muslim. In kelurahan Mambi the majority is Muslim; all the villages west of Mambi village are predominantly Muslim. In desa Meha1aan there seems to be a more equal distribution of Muslims and Christians, possibly with a majority of Christians. But the villages in the southern part of the desa are predominantly Muslim. Desa Matangnga was also reported to be Muslim, Sumarorong is predominantly Christian, while the areas visited in kecamatan Wonomulyo to the south are all Muslim. The church in the PUS area is part of the Gereja Toraja-:-Mamasa denomination. The Christian community, therefore, looks to Mamas a for spiritual direction, while the Muslim community looks to the coastal area. Health Generally health seems to be good throughout the area. In addition to the clinic in Mambi there were reported to be health centers in the villages of Mehalaan, Matangnga, and Sumarorong. In kecamatan Wonomulyo the people go to the town of Wonomulyo for health care. Thus, there are health facilities within a few hours reach for everybody in the area we visited. Co unication The main markets of the central and southern parts of kecamatan Mambi are located in the villages of Mambi and Galung-Galung. So there is frequent travel between these and the other villages of the area. The traffic is especially heavy to 38 PUS2 and from Mambi, which is also the governmental and educational center of the area. People also seem to travel to Malabo, Mamasa , and the coastal towns quite frequently. People in Matangnga and the southern part of Mehalaan especially seem to be more oriented towards the coastal area. They usually travel straight south to Tapango, while those further north prefer to follow the road via Malabo. The road from Malabo to Mambi was still 1n very poor condition. Unless extensive repair work is undertaken, it is not even recommended for motorcycles, although with much effort it is passable most of the way. The trail from Sika via Leko and Mehalaan to Keppe was considerably better suited for motorcycles than the main road to Mambi, so if one is able to make it to Sika, that route is a better alternative. In general the trails in the area are best suited for walking or horse riding, which is how most people travel there. Some of the trails to the more remote villages are not even recommended for horse riding. All the villages visited in kecamatan Wonomulyo are accessible by road good enough even for cars. A new road, eventually intended to reach Mambi, 1S presently under construction northwards from Po1liwa. There are also plans for improving the road between Malabo and Mambi. Construction of an airstrip in the central or southern parts of kecamatan Mambi would be very difficult due to its ruggedness. The only areas sufficiently level would be the rice fields around the villages of Mambi and Keppe, but destroying the fields for such a purpose would be unthinkable. Also these areas are within a day's walk or horse ride from the vehicular road at Malabo. Building an airstrip would also be very unwise in view of the fact that a new road into the area and an improvement of the present one is being planned. CORCLUSIOR Although there are still several unsolved problems and uncertainties with regard to the language situation in the PUS area, we nevertheless feel we have attained our major goal for the survey and the ensuing analysis, i.e. to reach an overall picture of the language situation in pus. Disregarding the Tabang area and fringe villages such as Periangan and Kondo, we have concluded that two languages are spoken in the area traditionally named Pitu U1unna Salu. We have chosen to call these languages Pitu Ulunna Salu and Aralle-Tabulahan. We have suggested eight dialects within the PUS language and three within the Ara11e-Tabulahan language. Furthermore, we have shown that those languages are distinct from those of the adjacent areas to the north (represented by the PUS2 39 Periangan wordlist), to the east (Kamasa), to the south (Pattae', Pannei, and the language represented by the village of Kondo) and from the coastal Mandar dialects. Finally we have suggested a PUS subfamily of languages including PUS, Aralle-Tabulahan, Pannei, and probably one language in the area between these and Mandar. Within the PUS area, therefore, we see an immediate need for two linguistic field programs. An additional program is needed for the Pannei language to the south and probably at least one more in the area between PUS and Mandar. References Esser, S.J., 1938. Talen. In Atlas van Tropisch Nederland. Amsterdam: Koninklijk Nederlandsch Aardrijkskundig Genootschap. Friberg, Timothy. UNHAS-SIL sociolinguistic Survey: Kabupaten Polewali-Mamasa, Northern Section. In this volume. Grimes, Charles E. and Barbara D. Grimes, 1984. Languages of South Sulawesi. To Appear. Mandadung, Arianus, 1982. Mamas a (Kondosapata' Waisapalelean) Dalam Informasi -Sejarah, -Budaya, -Pariwisata. Mills, R.F., 1975. Proto-South Sulawesi and Proto-Austronesian Phonology. Dissertation. University of Michigan. Pelenkahu, R.A., et. al., eds., 1974. Peta Bahasa Sulawesi Selatan. Ujung Pandang: Lembaga Bahasa Nasional Cabang III. Salzner, Richard, 1960. Sprachenatlas des Indopazifischen Raumes. Wiesbaden: Otto Harassowitz. Samar, Abd. Azis and Arianus Mandadung, 1979. Ungkapan Sejarah . dan Budaya di Kabupaten Polewali-Mamasa, Sul-Sel. -Seri: "A" = (Daerah Kondosapata'!Mamasa). 40 PUS2 UBBAS-SIL Sociolinguistic Survey: Seko Area (Kabupaten Luwu) Thomas V. Laskowske Kathryn B. Laskowske TABU OF COBrUTS Page Introduction 42 Survey Account 42 Government, Population 43 Background of What Is Meant by "Seko" 43 Language Analysis and Results 44 Conclusions about Language Relationships 46 The Seko People 48 Conclusion 50 Footnotes 51 References 51 Figures 1. Villages and Government Boundaries in Seko 52 2. Chaining Relationships in the Seko Language Family 48 3. Map of Languages in Seko 53 Tables 1. Government and Population of Seko 54 2. Percentage of Lexical Similarity 55 3. Summary of Languages Spoken in Seko 56 4. Number of Speakers of Seko Languages 57 5. Lexical Similarity Summary 57 S~KO 41 IBTRODUCTlOR The survey of the la~guages of the Seko area was conducted as a follo~up to the survey by Grimes and Grimes of the languages of south Sulawesi (to appear) in 1982-83. Our purpose was to determine tentative language and dialect variations and their boundaries in the area known as Seko by comparing lexical similarity from wordlists taken in the Seko villages. We also sought information about sociolinguistic factors to help us know how to conduct further language study in Seko. The survey' team visited all twenty-three villages in the Seko area. Wordlists and sociolinguistic data were obtained in most of those villages, exceptions being villages where we had no reason to suspect that the linguistic and sociolinguistic data was different from"other villages. The participants of the survey team who collected and analyzed the data were Thomas V. Laskowske and Kathryn B. Laskowske. SODRI ACCOUNT The survey team left Ujung Pandang on October 13, 1984 headed for Pa1opo. On October 15 and 16 we met the Bupati (District Government Head) H.A. Mubara nappi, the DanDim (District Military Commander), the Kapolres (District Chief of Police), and the Kepala Cabang Dinas Perkebunan (Branch Director of the Department of Plantations) at their respective offices in Pa10po, capital of the Luwu District. Letters were obtained for the Camat (Subdistrict Government Head) of Limbong and for the subdistrict police. (The purpose of meeting with the Plantations Director was not part of our language survey, but was done as a service to BAPPEDA (Regional Development Planning Board) and Dinas Perkebunan to investigate the suitability of Seko for growing tea). We then went to Sabbang where we arranged for most of our things to be carried by horse for the four-day walk into Seko (see map, Figure 1 in Appendix). The morning of October 17, we drove our motorcycle to Kanandede and reported to the subdistrict police. We stored our motorcycle there, at the home of a local official, although in hindsight we decided it would not have been very difficult to drive to Salutallang. We continued our journey walking to Komba where we reported to the Danramil (subdistrict military commander). We then continued on to Kawalean for the night. In Sa1utallang, we met with S. Palu1lungan, the Wakil Camat (Deputy Head of the subdistrict government), who wrote letters to the kepala desas (Heads of local government) of the five desas we planned to visit. On October 19, we left Saluta11ang with a Seko Padang man and later joined up with his five horses and his 42 SEKO family. We arrived in Eno the evening of the second day and began to gather linguistic and sociolinguistic data there. Eno is centrally located in Seko Padang, so we used that village as our home while travelling to surrounding villages in Desa Padang Raya and Desa Padang Balua'. We gathered data in the following villages: Eno, Lodang, Bengke, Tanete, Parawa1eang, Bana and Singka10ng. While in Seko Padang, we also looked over a piece of land which is a good site for a small airstrip. The site is free of trees or bushes, has excellent overall clearance, and in generally very level, although some work would be required before it can be used. The only disadvantage is that it is slightly under 400 meters long. From Seko Padang, we continued on November 1 to the area called Seko Tengah, made up of the one desa Tana Makaleang. Of the five villages in this area, we stopped and gathered data ~n four: Amballong, Pewaneang, Pokappaang and Hoyane. Finally, we went to Seko Lemo, comprised of the two desas, desa Tirobali and desa Malimongan. We passed through all the villages in these two desas, stopping and gathering data in the two kota desas (principal village of the desa): Rantedanga' and Kariango. From Seko Lemo we walked back out of the Seko area, reaching Salutallang on the second day where we reported at the Camat's office. We continued on that day, reporting in at Komba (Koramil) and Kanandede (police). Since leaving Kanandede on foot on October 17, we had walked 333 km (207 miles). We spent the night in Kanandede and returned the next day, the 11th of November, to Sabbang and Palopo. On the 12th we reported back in to the Bupati, DanDim, the police and the Kantor Perkebunan. That night we returned to Ujung Pandang, arriving there early on the morning of November 13, 1984, a month after leaving. GOVERRMERT, POPULATION Seko is located entirely in kecamatan Limbong, kabupaten Luwu, South Sulawesi. About 16 kilometers north of Seko is the Central Sulawesi border. All Seko is divided into three parts, consisting of five desas as listed in Table 1. (See Appendix) Rongkong is the area along the road going into Seko and is divided into two parts. Lower Rongkong (Bawah) extends from Sabbang to Buka, and Upper Rongkong (Atas) extends from Kanandede to the vicinity of Salutallang. BACICGROURD OF WHAT IS MEABT BY "SElCO" Seko is separated from Rongkong by a distance of about 50 km of uninhabited forest, as well as a mountain/watershed divide. All the rivers in Seko flow west into the Karama River of Mamuju, whereas all Rongkong rivers flow east to the Bay of Bone. The SEKO 43 • geographical isolation causes outsiders to perceive the area as a unit. Furthermore, during the time of the Dutch rule, the area was unified under an administrative entity and COllectively referred to as Seko. Previously, however, only the people in Seko Tengah referred to themselves or their language with the name "Seko", as they still do today. The term Seko is derived from a word which means sahabat or "close friend". The inhabitants of Seko Lemo are relatively recent immigrants from Upper Rongkong, and they speak the Rongkong language. The people living in Seko Padang prefer to call themselves 12 Padang, "the people of Padang". and their language Sua To Padang, "the language of the people of Padang". They don't seem to mind, however, their language being referred to as Bahasa Seko Padang or Bahasa Seko, following the tendency for groups in this part of the world to name their language after the geographical/political area in which they live. LANGUAGE ANALYSIS AND RESULTS The focus of our survey was the languages of the people living in Seko. Our main method for determining language differences and boundaries between languages was taking wordlists in representative villages, and then comparing those wordlists. To further understand the language situation, we took sociolinguistic questionnaires in various villages, and also asked informal questions as we stayed with the people or walked the trails with them. 204 of the 216 words on the wordlist were compared. 12 words were eliminated because we felt those words 1) were too difficult for the respondents to~ answer satisfactorily, usually because there was not a one-to-one lexical correspondence between their languages and Indonesian or 2) would bias the results because of lexical repetition. After comparisons were made, the percentages of lexical similarity were tabulated (see Table 2 in Appendix). We determined that two words were apparent cognates when we could readily reconstruct one word from the other based on generally accepted comparative reconstruction principles for phonetically similar segments. Thus we are recognizing genetic relationships without proposing the proto-forms. This is an attempt to approximate a diachronic approach, which we feel is valid, but we did not rigorously construct correspondence sets nor did we analyze for loan words, as would be done in an in-depth analysis for true cognates. We felt it was beyond the scope of this survey to attempt a historical reconstruction. The term "cognate" where used in the rest of this paper should be understood to mean apparent cognate, an initial approximation of true cognate percentages. 44 SEKO In some cases we made opposite dicisions than if we had followed a synchronic approach. For example, we called [bua] cognate with hoa "fruit" even though more than half of the phonetic segments are different, because b is in a regular correspondence set weth h, and u and 0 are ~both back vowels. Again, ono is cognate~ with -unun "six" because in the first language there are no [n]s at the ends of words, and because the o's and u's are back vowels. On the other hand, we did not count [bo?u and-ba?ru] "new" as cognate even though more than half of the segments are the same. We may have counted them as cognate if there were no other evidence available, but someone from another village gave both hou and bak aru for this word. Thus, hou is cognate with [bo?u] because of regular b/h correspondence sets between these languages; [ba"ru] is in correspondence with bak aru through the alternation of ?: with k and the addition of an a Even though our methodS-were somewhat different from those of Grimes and Grimes, When we compared our cognate dicisions with theirs on the same wordlists, the average of the six comparisons was exactly the same. The average variaton was 1.3% with the greatest variation at only 2%. We conclude that our methods for making cognate decisions produced very similar results. According to the threshold percentages used by Grimes and Grimes in their work, two wordlists are in the same group if the percentages are as follows: 80-100% same language 75-80 % same subfamily 60-75 % same family 45-60 % same stock 25-45 % same superstock These values are useful for prelimenary classification based on lexical similarity, only until in-depth analyses are made of the phonologies, grammars, and lexicons. In particular, intelligibility testing is needed to differentiate languages. Nevertheless, we appreciate the value of measuring lexical similarity, even for initially differentiating dialects. Theoretically, if two people speak the same dialect, they know and use the same phonology, grammar, and lexicon. Therefore, one would expect 100% similiraty between two wordlists of the same dialect. However, when we compared wordlists berween speakers of the same dialect (according to our sociolinguistic survey) in Seko Padang, the lexical similarity ranged from 91-98% with an average of 95%. A discussion of the factors which cause the percentages to be less than 100% is beyond the scope of this paper. However, we feel that the percentage will go no lower than the mid-nineties between two wordlists of the same dialect provided that the field linguist is experienced and that the respondent has a ready SEKO 45 knowledge of both his own language and the elicitation language. Where a respondent gives answers reflecting an unusually high number of Indonesian borrowings or misunderstandings, he may cause the percentage of lexical similarity to be further lowered 3-5 percent. (We had two such wordlists, Parawaleang and Kariango.) In order to allow for this possibility, we think percentages from 88-100% should intially be considered one dialect, with 88-92% being a gray~ area between same versus different dialects. Less than 88% similarity seems to represent truly different lexicons. The field linguist needs to distinguish dialects because of the possible need for adapting written materials for multiple dialects. Therefore, as a preliminary step, we suggest this guideline for differentiating dialects based on apparent cognate percentages: 90-100% one language, one dialect 81-90 % one language, two dialects with 88-92% as a gray area. Values above 90% do not prove that there are not two dialects, however. Significant differences in the grammars or phonologies should be taken into account when discovered. CONCLUSIOIIS ABOUT LAIIGUAGE RELATIONSHIPS Our results confirmed the results of Grimes and Grimes' language survey of South Sulawesi as follows: 1) Seko (the language) should not be included in the northern South Sulawesi Family, but should be included in the South Sulawesi Stock. Shared-cognate range 49-59%. Average 55%. 2) The Rampi language should be excluded from the South Sulawesi Stock. Range 34-41%, Average 38%. 3) Residents of Seko Lemo speak the Rongkong language. Range 85-90%, Average 88% (with the Rongkong Atas dialect). 4) Kalumpang, Rongkong, and Toraja should all be considered different languages within the same subfamily. Range Average Rongkong with Kalumpang 73-81% 77% Rongkong with Toraja (Grimes) 74-78% 77% Toraja (Grimes)~with Kalumpang 72-75% 74% 5) Our wordlist from Seko Tengah and Rongkong Atas agreed with the Grimes and Grimes' lists for those locations to the following degrees: 46 SEKO Range Average Seko Tengah 93-99% 96% Rongkong atas 96% 96% 6) Comparison of opposite ends of the Kalumpang area show 81% similarity. we affirm that the Kalumpang language could possibly extend from the Mamuju-Luwu border (Bau wordlist) into Kecamatan Mambi in Polmas. (Makki wordlist taken on the UNHAS-SIL PUS survey.) However, because 81% is only marginally the same language, there is a definite need for further survey and intelligibility testing. Our comparisons also revealed the following new information: 1) The Seko Family is composed of not -one, but three languages. Seko Padang was previously thought to be a dialect in relation to Seko Tengah, and Panasuan was unclassified. Range Average Seko Padang - Seko Tengah 66-75% 71% Panasuan - Seko Tengah 66-70% 67% Panasuan - Seko Padang 62-64% 63% We don't know whether there has ever been reference to a "Padang" language in Sulawesi. In the literature, reference has been made to only one Seko language, at best with mention of two dialects, as in the Grimes and Grimes' South Sulawesi survey. We consider the terminology "the Seko language" too general and therefore potentially confusing. We thus propose that the two languages should be distinguished by using the terms "Seko Tengah" and "Seko Padang". 2) Among South Sulawesi languages, Panasuan relates most closely to the Seko Tengah and Seko Padang languages (see percentages above; also see footnote **). Chaining relationships among languages in the Seko Family exist, but are not pronounced (Figure 2). SEKO 47 60% 70% 80% 90% Wono Lodang Seko Tengah Panasuan Seko Padang Figure 2. Chaining Relationships in the Seko Language FamiIy After the Seko languages, Panasuan relates most closely with Rongkong at 58% lexical similarity and Kalumpang at 53%, putting it in the same stock with these languages - the South Sulawesi Stock. 3) The Bana language is called To Bau by the speakers of the language and compares with Uma (Pipikoro) at 95% (compared with an Uma wordlist obtained from Michael Martens, December, 1984). The people themselves claim to originate from Kantewu (central Uma area), and we were told that they went to live in the forest when the Dutch came, to avoid being governed by them. Thus, it appears that Bana is not a separate language, but is the Uma language. 4) We compared the wordlist from Singkalong in desa Padang Balua' with a wordlist taken in 1983 in Onondoa, the central and largest village in the Rampi area (UNHAS-SIL Rampi survey). Lexical similarity was 92%, so we conclude that it is indeed the Rampi language. 5) People from Bau claim to speak Te'da, the Karama dialect of the Kalumpang language. Although people perceive the Karama dialect as a separate dialect, our data indicates that there is a very close relationship to the Kalumpang dialect (91%). When comparing both of these, however, to a Makki wordlist taken on a previous UNHAS-SIL survey (Dec 1983) in Kabupaten Polmas, the similarity was only 82%, indicating that this Makki word list represents a definitely separate dialect, not just an alternate name for Ka1umpang. 6) The wordlist taken in Lodang ranged 84-91%, average 88%, similar to wordlists taken in the rest of Seko Padang. Lodang should tentatively be considered a separate dialect of the Seko Padang language. TIlE SEKO PEOPLE Economy - Subsistence farming is the livelihood of almost everyone in Seko, as it has been for centuries. Basic foods are grown, gathered or produced locally, including rice, vegetables, 48 SEKO meat, fish, coffee and cane sugar. The Seko economy is only marginally based on money. Instead, the barter system iS used, both among themselves and with traders who come through the area with horseloads of goods. Cloth, kitchenware, flour, white sugar, etc., can be obtained by trading coffee beans. Sometimes the Seko people will load their own horses with coffee, and make the 4-to-5-day trip to Sabbang or Palopo and sell it for cash. Because markets are so difficult to reach, it is not profitable for Seko people to increase food production, even though the land is fertile and could support more. Seko's most distinctive geographical feature is its isolation. The absence of a serviceable road connecting Seko with coastal areas is the foremost problem mentioned by local leaders. Some years ago there was a road to Seko, but due to lack of maintenance it has become impassable except to horses and foot traffic. It may be several years before existing roads will be improved or new roads built in Kecamatan Limbong. The work will probably be done in stages, with uncertainty as to when it might finally reach Seko. The altitude of the villages in Seko ranges from about 900 to 1500 meters. This has proved to be suitable for coffee production. It is said that 1100-1500 meters is also suitable for growing tea. The land in Seko Padang is broad and expansive, while in Seko Tengah and Seko Lema it is steep with very little space for expanding their rice paddies. Most of the rice grown in Seko Tengah comes from dry-land cultivation. On the other hand, Seko Padang has much land suitable for paddy rice, some of which has been used as such in the past, but which now lies fallow, because owners have moved to Central Sulawesi. (It is said that they are unwilling to move back to Seko, because the isolation prevents them from education and marketing opportunities that ~they presently enjoy in Central Sulawesi.) In the Seko region, Seko Padang seems to have the greatest potential for growth both in population and economy. Education Many villages have their own primary school. However, in the whole Seko area of five desas, including 23 villages, there is presently only one SMP (Junior High School) in Pawaneang, Desa Tana Makaleang. This SMP reflects the desire of the people for education, as we were told that it iS a volunteer-type school, run by the local people rather than by the government. There is no SMA (High School) in the area. The closest SMA is in Sabbang, four days' walk from Seko, a distance of about 125 km from most villages. A road would enable more to attend this and other schools. Regardless of limited education opportunities, most of the Seko people we met could speak Indonesian, exceptions being children up to age 10-12 and some elderly adults. SEKO 49 Religion - All of the people in the Seko area were reported to be either Christian or Islam, with a greater percentage of Christians in each language group. Seko Lemo, speaking the Rongkong language, is unique in that it is about 99% Christian, while speakers of the same language in the Rongkong Atas area are about 80% Islam. Health Facilities - We did not hear of any health facilities in Seko. The closest, a Balai Kesehatan (medical clinic), is two days' walk away, in Salutallang. Here again, a road is a felt need. (After returning to Ujung Pandang, however, we met a young man who said that his sister has been running a Balai Kesehatan in Pokappaang for 10 years.) CONCLUSION By completing this survey and subsequent analysis, we feel we met our objectives of mapping the languages in Seko based on lexical similarity. We have determined that six languages are spoken in Seko, but four of them are centered outside of Seko. Only Seko Tengah and Seko Padang are spoken primarily in the Seko area. These two are best considered as separate languages, rather than as dialects of the same language. Furthermore, we have proposed to classify Panasuan as part of the Seko Family. There is a need for further linguistic field work in all three of these languages. 50 SEKO FOOTNOTES ** There is also a folktale told in Seko Tengah about the common origin of the Panasuan and the Seko people which strengthens the probability of an ethnic relationship between them. REFERERCES Grimes, C. E., and B. J. Grimes. To Appear. Languages of South Sulawesi. Pacific Linguistics, Series D. Canberra, Australia. Laskowske, Thomas V. In this volume. UNHAS-SIL sociolinguistics Survey: Rampi Area (Kabupaten Luwu). Friberg, Timothy. In his volume. UNHAS-SIL Sociolinguistic Survey, Kabupaten Po1ewa1i-Mamasa, Northern~Section. SEKO 51 APPENDIX- Maps and Tables Figure 1 - Villages and Government Boundaries in Seko 52 SEKO Figure 3 - Mapof Languages in Seko SEKO 53 Table 1 CovernmeAtaDd Population of Seko Area Desa Popu1ation(1) Village(2) Popu1ation(3) Seko Padang 1843 Eno 426 Padang Balua Tanete 455 Bana (4) 120 Kalammio 200 Parawaleang 237 Kampung Baru 100 Tanete Huko 50 Singkalong 457 Padang 881 Lodang 527 Raya Bengke 200 Busak 200 Seko Tana 2288 Pokappaang 700 Tengah Maka1eang Pewaneang 500 Hoyane 400 Poak-Poak 370 Amba110ng 532 Seko Tirobali 1721 Rantedanga 685 Lemo Bua Kayu(5) - Kampung Baru - Seppulung - Beroppa' - Malimongan 921 Kariango 500 Se'pon 420 Total 7654 (1) Desa population figures were obtained from the Camat's office. (2) The first village listed for each desa is the Kota Desa. (3) Village population figures were generally obtained on site from local leaders. (4) Not included in this figure are 63 Bana people who, at the time, were not living in the Bana village. (5) A hypen (-) means that we did not obtain a population estimate from anyone while we were in that area. 54 SEKO Table 2. Percentage of Lexical Similarity SEKO 55 List of Abbreviations for Table 2 SNG Singkalong STG Seko tengah-Grimes BAN Bana HOY Hoyane PAR Parawaleang AMB Amba110ng TAN Tanete PAN Panasuan EN1 Eno. man from outside Seko Mak Makki - PUS Survey EN2 Eno. woman from outside Seko KAL Ka1umpang - Grimes EN3 Eno. man from Seko BAU Bau EN4 Eno. woman from Seko KAR Kariango BEN Bengke BAN Rantedanga' LOD Lodang RKG Rongkong Atas-Grimes PEW Pewaneang RKK Kawa1ean POK Pokappaang TOR Toraja - Grimes Table 3 - SUMmary of Languages Spoken in Seko Language Name Village Where Spoken Seko Padang Padang Ba1ua' (in Eno, Tanete, Ka1ammio and Parawa1eang) Padang Raya (in Lodong and Bengke) Uma Padang Balua' (only in Bana) Rampi Padang Balua' (only in Singka1ong) Seko Tengah Tana Maka1eang (in all the villages) Rongkong Tiroba1i (in Rantedanga',Kampung Baru, Seppu1ung and Beroppa') Ma1imongan (in both villages) Ka1umpang Tirobali (only in Bua Kayu) (Dialect Te'da) 56 SEKO Table 4 - Humber of Speakers of the Three Main Languages in Seko Language Name Number of Speakers Seko Padang About 2100 living in Seko. There are reported to be one to two times that many Seko Padang speakers in Central Sulawesi, people who moved there in the early 1950s and have not returned. Seko Tengah 2502. Rongkong 2992 living in the Seko area. There are an additional 3039 speakers of this linguage living in the Rongkong Atas area, south of Seko. This totals 6031 speakers of Rongkong, not counting the Rongkong Bawah dialect. Table 5 - Lexical Similarity Summary Pamona Rampi Family Central Sulawesi Kaili Stock 44 Uma Family 40 38 Seko Padang Seko 39 39 71 Seko Tengah Family South Sulawesi Stock 37 39 63 67 ~ Panasuan 36 37 51 54 58 ~ Kalumpang Northern ~ ~ South Sulawesi 36 39 53 56 58 77 Rongkong Family Toraja 34 38 47 50 53 74 77 ~ Toraja Sub-Family SEKO 57 page 58 [blank] UNHAS-SIL SOCIOLINGUISTIC SURVEY KABUPATEN POLEWALI HAMASA, SOUTHWESTERN SECTION, AIm KABUPATEN MAJENE Kare J. Stromme and Kari Valkama TABLE OF CONTENTS o. INTRODUCTION 60 1. NON-LINGUISTIC INFORMATION 68 2. SURVEY ITINERARY 76 3. METHODOLOGY 77 4. RESULTS 83 5. RESIDUES - FURTHER RESEARCH 91 6. CONCLUSION 91 BIBLIOGRAPHY 94 TABLES: 1: Government and Population of Kabupaten Polewali Mamasa, Southwestern Section, and Kabupaten Majene 69 2: Reduced Matrix of Lexical Similarity Percentages and Language Grouping 82 3: Languages of Kabupaten Polewali Mamasa, Southwestern Section, and Kabupaten Majene with Estimated Number of Speakers 86 MAPS: 1: Kab. Polewali Mamasa, Southwestern Section, 72 and Kab. Majene 2: Languages of Kab. Polewa1i Mamasa, Southwestern Section, and Kab. Majene 84 MANDAR 59 o. IIIf'1dJDOCTI" 0.1. Previous Work Several language surveys have previously been undertaken in the area covered by our survey, i.e. the southwestern section of kabupaten (district) Polewa1i Mamasa and kabupaten Majene, and descriptions of the linguistic situation of the area are found in a number of books, theses, articles, and reports written throughout this century. Being important background material for our survey, we will try to summarize the views expressed in the most important of these works with regard to language and dialect divisions and classifications for this area. Adriani and Eruijt 1914 The first significant description of the area is found in Adriani and Kruijt's De Bare'e-Sprekende Toradja's ~ Midden-Celebes, Derde Deel -r1914). In this work Adriani and Kruijt suggest a Mandar group consisting of the languages of Mandar and Mamoedjoe. All the area we surveyed, belongs to their Mandar language area, which they divide up into five different dialects: Tjenrana (located in present kecamatans (kecamatan = subdistrict) Sendana and Ma1unda), Madjene (located in present kecamatans Banggae and Pamboang), Ba1angnipa (located in present kecamatans Tinambung, Tutallu, and Wonomu1yo), Tjampalagian (located in present kecamatan Campalagian), and Binoeang (located in present kecamatan Polewali). As to the extent of these dialects inland, Adriani and Kruijt are not too clear. On their Language Map of Celebes the boundary between the Mandar and the Sa'dan language group is drawn as far east as the Masupu river in kabupaten Tanah Toraja. In the north, Rante Boe1awang and Sa10 Tabang are suggested to be dialects of the Mamuju language, but they do not seem to know exactly where these places are located. In a postscript to their book and in later articles by them in Encyc10paedie van Neder1andsch-Indie, they made quite extensive revisions of these boundaries, however. Here the Sa'dan language is said to extend as far west as to the present border between kabupatens Majene and Polewa1i Mamasa (apparently also including the Ulumandak area of kabupaten Majene). In the south the boundary between Sa'dan and Mandar is drawn almost straight eastwards from the southern part of present kecamatan Tutallu. Thus, the whole of the Pitu Ulunna Salu (PUS) area falls within their Sa'dan language group. Apart from a Tjenrana wordlist which they collected themselves, Adriani and Kruijt's conclusions are all based on language material of limited reliability collected by Dutch 'civil and military personnel. They obviously did not do extensive trav~lling in the area themselves. Their conclusions, therefore, must be considered tentative. 60 MANDAR van der Veen 1929 In his article "Nota betreffende de grenzen van de Sa'dansche taalgroep en het haar anverwante taalgebied" (1929), H. van der Veen also gives a lot of valuable information about Sa'dan's neighbouring areas, including the area covered by our survey. Whether his descriptions are based on personal surveys in the area or on secondary information, is not clearly stated, but the many details provided and the large extent of agreement with our and previous UNHAS-SIL field research in the area would seem to indicate that the former is the case. The most significant difference between van der Veen and Adriani and Kruijt is that the former separates the PUS area from the Sa'dan language, thus establishing a PUS language. The boundaries given for this language, roughly follow those of Adriani and Kruijt for the Sa'dan language in the west and the south (after their revision), but in the east the boundary is drawn west of the Mamasa river. As his boundary between PUS and Mandar are of particular interest to this report, we will try to account for this in more detail, using present names for administrative units and villages. In the southeast it starts in the northern part of desa Matakali (immediately to the east of desa Palitakan) at the boundary between kecamatans Polewali and Wonomulyo, approximately following the southern boundary of desas Palitakan and Rappang to the west, crossing the Mapilli river to the east of Lena village, and from there continuing straight west approximately to the boundary of kabupaten Majene (following a course somewhat to the north of that indicated by Adriani and Kruijt in their postscript). Specifically the villages of Labasang, Tapango, Bussu, Landi, Batu, and Dakka of desas Matakali, Tapango, and Palitakan in kecamatan Wonomulyo, as well as kelurahan (administrative unit corresponding to desa but having a higher status) Taramanu and desas Ambopadang and Tubbi of kecamatan Tutallu are said to belong to the PUS language group. In the west, van der Veen's boundary between Mandar and PUS approximately follows the kabupaten boundary and the western boundary of desa Ulumandak up to the boundary of kabupaten Mamuju. Specifically van der Veen says that the PUS language is spoken in the mountain areas of what is now called kabupaten Majene and mentions in that connection the Pupenga people (in the interior of present kecamatan Malunda) and the Oeloemanda and Majamba districts of Tjenrana. Oeloemanda probably roughly corresponds to present desa Ulumandak, while Majamba must correspond to the interior part of one of the other desas of kecamatan Sendana or Malunda. Most of the inhabitants of what is now kecamatan Tapalang are also said to speak the PUS language. As for the Mandar language, van der Veen seems to follow the dialect divisions of Adriani and Kruijt, although he does not MANDAR 61 mention anything about the Campalagian dialect. As did Adriani and Kruijt, he combines the Mandar and Mamuju languages in a Mandar group. The boundary between the Mandar and Mamuju languages van der Veen seems to draw at the present kabupaten boundary, as do Adriani and Kruijt, although they in one case also seem to include Tapalang with Mandar. Esser 1938 S.J. Esser's classification and language map in Atlas van Tropisch Nederland (1938) basically follow van d~r Veen's suggestions for the area concerned, separating PUS from Sa'dan. But Esser does not mention Mamuju as a separate language of a larger Mandar group. For the whole coastal area between Polewali and Karossa in Mamuju he uses the term "Mandarsche dialecten". Salzner 1960 In R. Salzner's ~rachenatlas des Indopazifischen Raumes (1960) PUS is again included under Sa'dan, although now as a separate language of that larger group. Like his predecessors Salzner suggests a Mandar group consisting of the languages Mandar and Mamuju. The Mandar language he divides up in the five dialects suggested by Adriani and Kruijt. The Mandar and Sa'dan groups are again classified under the yet larger South Sulawesi group. The boundaries on Salzner's map are somewhat different from those of the previous researchers, though. The dividing line between PUS and Mandar, for instance, Salzner draws more to the interior than Adriani and Kruijt and van der Veen do, both in the south and the west. Tubbi is described as a dialect of PUS but on the map it is located too far to the south, as a separate enclave within the Mandar language. The Balanipa dialect of Mandar is located along the coast including all of kecamatan Polewali and most of Wonomulyo, while the Binuang dialect is located to the north, between Balanipa and the PUS language, extending far into kecamatans Sumarorong and Wonomulyo. The Majene dialect is shown to include the present kecamatans Pamboang, Banggae, Tinambung, and Tutallu, while the Tjenrana dialect includes kecamatans Sendana and Malunda. At least the locations of the Tubbi, Balanipa, and Binuang dialects reveal some lack of geographical knowledge of the area, but when covering such large areas as Salzner does, some inaccuracies are probably unavoidable. Pelenkahu 1967 In his dissertation "Proto-South Sulawesi and Proto-Austronesian Phonology" (1975) R.F. Mills refers to data obtained from R.A. Pelenkahu regarding the Mandar language. Some of this data is probably presented in Pelenkahu's thesis "Gambaran 62 MAN DAR Sepintas Lalu Tentang Dialek-Dialek Mandar" (1967), but as this work unfortunately is no longer available to the public, we can only comment on what Mills cites. The only thing of interest in this connection is that Pelenkahu is said to talk about a dialect of Mandar called Toda-Todang, which according to his map is spoken in the north, inland from Sendana. Interestingly, in Peta Bahasa Sulawesi Selatan (~ Petunjuk) (1974), of which Pelenkihu was one of the editors, Todatodang is mentioned as one of the variants of the Balanipa dialect. As there is a desa called Todang-Todang within the Balanipa dialect area (in the northeastern corner of kecamatan Tinambung), we assume that this is where this variant is spoken. The location cited by Mills seems to be further north, but that might well be an inaccuracy. Pelenkahu et ale 1974 Ceds.) The above-mentioned Peta Bahasa Sulawesi Selatan (Buku Petunjuk) (hereafter: Peta--Bahasa) , like many of the ear~ works, suggests a larger Mandar language group. This is then divided up into the subgroups of ~andar, PUS, Padang-Mamuju, and Botteng-Tappalang. Compared with the previous works, the area covered by this larger Mandar group, corresponds approximately to the area covered by van der Veen's Mandar group and his PUS language, and that of Esser's "Mandarsche dialecten" and "pitoe-oeloenna-saloesch". When including PUS in the Mandar group, the works of most of the previous researchers are disregarded. We actually have to go all the way back to Adriani and Kruijt's earlier writings to find anyone drawing the boundary of the Mandar group as far east as Peta Bahasa. Peta Bahasa's subdivisions Botteng-Tappalang and Padang-Mamuju are also new, but these areas are outside the area covered by our survey. The Mandar subgroup is said to consist of the dialects Balanipa, Majene, Pamboang, Sendana, and Awo-Sumakuyu. We note here that the dialects of Binuang and Campalagian, mentioned by all previous researchers attempting a dialect division of Mandar, are not included. About the former nothing is mentioned at all, while the latter (also called Tallumpanuae) is classified as a dialect of the Bugis language. With regard to Binuang we should mention, though, that in the report describing the field work on which the Peta Bahasa is based, it is actually stated that no dialect by this name was found in the Binuang area, but rather Bugis Pattae. Interestingly, Salzner mentions Tae as an alternate name for the Binuang dialect and Grimes and Grimes in their prepublication copy of Languages £i South Sulawesi list the separate language of Pattael (mentioning Binuang in parenthesis) for this area, which they have found not to relate closely either to Mandar or Bugis. The Peta Bahasa, therefore, seems to be right in eliminating the Binuang dialect from Mandar. Nothing is MANDAR 63 mentioned about Pattae' in the Peta Bahasa itself, however. On its map this area is assigned to the Sawitto dialect of Bugis. We also note that two new dialects of Mandar have been added, Pamboang and Awo-Sumakuyu. All that is said about the latter is that it is located in kecamatan Malunda, while the Pamboang dialect is said to be located on the coast of kabupaten Majene in the area between its southernmost point (Tanjung Mandar) and Sirindu (in the northernmost part of kecamatan Pamboang), consisting of the Luwaor-Bababulo and Pamboang or Adolang variants. The Balanipa dialect is said to consist of variants such as Lapeo, Pembusuang, Karama, Napo-Tinambung, Tandung, and Todatodang (apparently not an exhaustive listing) and to be located south of the PUS subgroup from the lower section of the Mandar river in the west to near kecamatan Polewali in the east. Balanipa speakers are also said to be found within kecamatan Polewali and the Tallumpanuae (Campalagian) speaking area, as well as at Ujung Lero, just outside the Pare-Pare harbour. Peta Bahasa's boundary between this dialect and PUS closely follows that given by van der Veen east of the Mapilli river, although, as rightly noted on Peta Bahasa's map, Bugis and Javanese LRmigrants have to a . large extent taken over on its southern side in that area. West of the Mapilli river Peta Bahasa draws the boundary further north than van der Veen does, apparently including part of desa Ambopadang, and most of kelurahan Taramanu (up to the village of Taramanu) and desa Pao-Pao with the Mandar language. The Majene dialect is said to be located around the town of Majene between the Mandar river and Tanjung Mandar, extending a little inland. It is said to consist of, among others, the Baruga, Tande, Labuang, Saleppa, Pengaliali, and Camba variants, the latter four being spoken within the town of Majene. All that is said about the Sendana dialect, is that it consists of several variants not yet identified and that it is spoken in kecamatan Sendana and several places in kecamatan Malunda. As for Peta Bahasa's boundary between this dialect and PUS, it appears from the map that it basically follows the kabupaten boundary up to the point where this turns east. From there it continues straight north, turning westwards at the Malun~a river, reaching the coast at Uluserang (near the kabupaten boundary on the Mamuju side). The interior parts of kecamatan Malunda, including all of desa Ulumandak, are thus included under the PUS subgroup. Compared with van der Veen, Peta Bahasa draws the PUS boundary somewhat further west here. 64 MANDAR Hilla 1975 Apart from some data of his own on the Balanipa and Majene dialects, R.F. Mills (1975) bases his discussion of Mandar on Adriani and Kruijt (1914), van der Veen (1929), and Pelenkahu (1967). (Peta Bahasa was probably not yet published when Mills wrote his thesis.) He therefore does not bring any new information regarding the language and dialect configuration of the area. It is clear that Mills follows van der Veen's separation of the Mandar and PUS languages. Regarding Mamuju he says that is ~s questionable whether such a language exists at all, tentatively concluding that it is basically PUS influenced by Mandar and other South Sulawesi languages. He admits, though, that the language situation of the Mandar area is still somewhat unclear, saying that lithemost pressing need is a thorough dialect survey, to determine as far as possible the exact boundaries and isoglosses." Gru.ea and Gru.es (to appear) Grimes and Grimes (to appear) for the most part base their analysis of the area on a survey carried out by C.E. Grimes and F.B. Dawson in January 1983. Supported by a lexicostatistic comparison of wordlists collected on that survey (taking a relatedness of above 80% to indicate one language) they suggest Campalagian, Mandar, Mamuju, PUS, and Pattae' to be separate languages, the first belonging to the Bugis Family and the others to the Northern South Sulawesi Family. Further, the Mandar language is divided up into six dialects: Balanipa, Majene, Pamboang, Sendana, Malunda, and Awo' Sumakuyu, all except the last being supported by a wordlist in their lexicostatistic comparison. Grimes and Grimes do not make it clear, however, what criteria they have used to distinguish between dialects. Since their Balanipa and Majene wordlists, relating to each other as high as 95% in their comparison, have been taken to represent separate dialects, we conclude that they must also have used criteria other than lexical similarity. The only dialect suggested by Grimes and Grimes that has not been mentioned in previous works, is Malunda. The location of the various languages and dialects of the area as given by Grimes and Grimes, differs somewhat from that of previous works, particularly on their map. Their Campa1agian boundary, for instance, does not include the area of Buku within the Campalagian language, contrary to Peta Bahasa. As for the Balanipa dialect of Mandar, 'Grimes and Grim~s extend this further to the east than Peta Bahasa does, even beyond the town of Polewali, although Peta Bahasa admits that there are some Balanipa speakers in kecamatan Polewali as well. The boundary between Balanipa and the PUS language is drawn further north in kecamatan Wonomulyo than in both Peta Bahasa and van der Veen, including desas Palitakan and Rappang within Balanipa. West of the Mapilli MANDAR 65 river, though, GrLmes and GrLmes' boundary follows that of Peta Bahasa quite closely. Their location of the Majene and Pamboang dialects is quite similar to that of Peta Bahasa, although the boundaries between Majene and Pamboang and between Pamboang and Sendana are drawn somewhat further to the north. Their location of the Sendana dialect also roughly corresponds with Peta Bahasa, although Peta Bahasa's boundary with PUS is drawn further east. As GrLmes and GrLmes have suggested a separate Malunda dialect, this must cover some of Peta Bahasats and earlier works' Sendana dialect area. Contrary to Peta Bahasa, Grimes and Grimes have located the Awo' Sumakuyu dialect on the peninsula in the northernmost part of kecamatan Sendana. Ulumandak has been included under PUS, as in many previous works, but with the new dialect name Ulunda. As for the boundary between the Mandar and Mamuju languages, Grimes and Grimes draws that somewhat further south than previous researchers, approximately at the Lambong village in kecamatan Malunda. Str•••••me 1985 The UNHAS-SIL survey of the west central section of kabupaten Polewali Mamasa carried out b~ T. Laskowske and K.J. Str~e in September-october 1984, also covered some of the fringes of our present area. Some of the findings from that survey as explained by Str,hnne in his report "UNHAS-SIL Sociolinguistic Survey: kabupaten Polewali Mamasa, West-Central Section" (in this volume), provided us with some helpful background information. Pannei, for instance, was found to be a separate language. In line with previous researchers who have included this area under the PUS language, Str~mme included Pannei in a PUS Subfamily along with the PUS and Aralle-Tabulahan languages. Thus, although the Dakka language was not encountered on that survey, the approxLmate northern boundary of the Mandar language in that area as suggested by many previous researchers, was confirmed. Th~ village of Kondo in the southernmost end of kecamatan Mambi was found to speak a language sepa~ate from both PUS, Pannei, and Mandar. Information obtained there indicated that this language at least also included the northern part of desa Bu10, desa Tubbi, and possibly also parts of desa Ulumandak. Str6mme further explained that at Galung in the western part of kelurahan Mambi, it was reported that the closest villages in desa Ulumandak, i.e. Pupenga and Urekang, spoke the same dialect, i.e. the Pattae' dialect of PUS. Under the section "Present Residues - Further Research" in his report, Str.mme summarized the situation as follows: "We are unsure where the language boundaries of the language represented by Kondo are, we do not know the precise extent of the Pannei language, and we are generally unsure about the linguistic 66 MANDAR situation in the area between PUS and Mandar (desa Tubbi and the northern part of desa Bulo in Polewali Mamasa and desa Ulumandak in Majene). A linguistic investigation of this area is needed to answer these questions." 0.2. Aim. of Survey and Report In the light of all the linguistic research that has previously been carried out with regard to kabupaten Majene and the southwestern section of kabupaten Polewali Mamasa, as described above, the basic aim of our survey was to investigate in more detail the sections of that area where previous researchers were clearly uncertain and/or have come to conflicting conclusions. This in particular seemed to involve kecamatan Tutallu, the northern part of kecamatan Campalagian, desas Buku and Palitakan in kecamatan Wonomulyo, as well as the northern part of kecamatan Sendana and kecamatan Malunda in kabupaten Majene. In other words we particularly wanted to ascertain the northern boundary of the Mandar language in kabupaten Polewali Mamas4, the existence and extent of a language between Mandar and PUS (as indicated by Laskowske and Str~ets Kondo wordlist), the status of Dakka and Buku, the extent of the Sendana dialect to the north, the existence and location of an Awo' Sumakuyu dialect of Mandar, the extent of the Malunda dialect, the status of the Ulumandak area, and the precise boundary between. the Mandar and Mamuju languages. Through the anatysis of the 26 wordlists and 10 sociolinguistic questionnaires obtained from the area, as well as of more informally gathered information, most of these questions have been answered. Recognizing the limitations of our method, lexical comparison of wordlists, we did not see it as the task of this survey to establish in greater detail the boundaries between the Balanipa, Majene, Pamboang, and Sendana dialects or to identify and document the alleged variants of each of these dialects. That task will require another type of survey and other methods. No previous researchers have commented on the language use patterns in the area. We therefore also saw it as an important aim of our survey to obtain information in that area 1n order to ascertain the vitality of the local languages. The main aim of this report, therefore, is to present, on the background of and in comparison with previous research, the findings of our survey and the following analysis with regard to the questions mentioned above. But first we will include various nonlinguistic background information on the area surveyed, some of which will be of some importance to an understanding of the linguistic situation and some just of general interest to anyone carrying out further research in the area. MANDAR SURVEY page 67 1. HOHLIIIGUISTIC IMFOlIM'ATIOII 1.1. Government and Population Our survey was carried out in two kabupatens, Polewali Mamasa and Majene. In Polewali Mamasa we visited kecamatans Wonomulyo, Campalagian, Tinambung, and Tutallu, while our work in Majene was carried out in kecamatans Sendana and Malunda. The boundaries of the various administrative units are shown on Map 1 and population figures in Table 1. 1.2. History , Archaeological discoveries made in the Kalumpang area of kabupaten Mamuju indicate that there was a civilization in the northwestern part of South Sulawesi as early as in the Stone Age. This early population of the area was later mixed with people moving in from the Southeast Asian continent in several waves, resulting in the present ethnological configuration. (Rahman 1984:37) The history of the area is usually dividied into three periods. In the first of these, the Tomakaka period, the area consisted of a number of small sociological groupings ruled by leaders called Tomakaka (3 one who is made the head). This period is viewed as rather primitive, with law and order almost nonexistent. (Rahman 1984:45, Kallo 1983:6, Sahur 1976:19-20) Then followed the Papuangang period, a transitional period between the rather chaotic Tomakaka period and the quite well developed Arajang or kingdom period. The Arajang period began with the formation of the Balanipa kingdom about 1500 A.D. The kings of this period all trace their origin back to the mythical figure Tumanurung, who, according to the legends, descended from heaven somewhere at the headwaters of the Sa'dang river. Seven kingdoms emerged along the coast of what is now the three kabupatens Polewali Mamasa, Majene, and Mamuju, i.e. Binuang, Balanipa, Majene, Pamboang, Sendana, Tapalang, and Mamuju, as well as seven kingdoms in the interior area. The second king of the Balanipa kingdom, Tomepayung, in the 16th century suceeded in uniting the seven coastal kingdoms in a federation called Pitu Babana Binanga (lit.s seven - mouth of - river) of which Balanipa was the leading power. Not long after this the seven interior kingdoms formed the pitu U1unna Sa1u (lit.- seven - head of - river) federation. These two federations then went together in a larger federation at the so-called Luyo conference. The intention was primarily to improve their defence against common enemies, pitu Babana Binanga guarding against enemies attacking from the sea and Pitu Ulunna Sa1u agaist enemies coming overland. Thus, there were both strong cultural and political ties between these two groups. It is not surprising, therfore, that one term came to be used for the whole area represented by them, i.e. Mandar. The 68 MANDAR Table 1: Government and Population of Kabupaten Polevali Ma.asa, Southwestern Section, and Iabupaten Majene Kabupaten Kecamatan Kelurahan Pop. Village Pop. Pop. or Desa represented by wordlist Polewali- Wonomulyo Sid6dadi 10529 Mamasa Sumberjo 8229 370520 76568 Ugi Baru 5491 Mapilli 7840 Rumpa 2323 Buku 2534 Buku Tumpiling 3864 Matakali 7082 Bumi Ayu 4173 Kebunsari 5303 Palitakan 4914 Dakka Tapango 4088 Tapango Rappang 5013 Bulo 5125 Pulliwa Bulo Karombang Campalagian Pappang 4093 Campalagian 59618 Bonde 4212 Parappe 2730 Panyampa 1892 Katumbangan 5733 Lampoko 6966 Lapeo 6496 Suruang 3316 Ongko 2955 Sumarang 4394 Tenggelang 5209 Baru 5960 Batupanga 5542 Batupanga MANDAR SURVEY page 69 Kabupaten Kecamatan Kelurahan Pop. Village Pop. Pop. or Desa represented by wordlist Tinambung Pambusuang 8399 56240 Sabag Subbik 5083 Tammangalle 3615 Balanipa 2313 Karama 5859 Tandung 3468 Lekopadis 2024 Galung Lombok 2357 Tandasura 2898 Lembang-Lembang4318 Limboro 2699 Samasundu 2394 Napo 3724 Napo Todang-Todang 2472 Tinambung 4617 Tinambung Tuta1lu Mombi ~834 19566 Al1u 4627 Petoosang Pao-Pao 2449 Pao-Pao Taramanu 2429 Lombang Taramanu Ambopadang 2180 Ambopadang Tubbi 2376 Rattepadang Tubbi Pirian Besoangin 2671 Besoangin 70 MANDAR Kabupaten Kecamatan Kelurahan Pop. Village Pop. Pop. or Desa represented by wordlist Majene Banggae Banggae 15914 Majene 123319 52828 Labuang 15956 Totoli 12318 Baruga 4914 Tande 3726 Pamboang Lalampanua 8360 Pamboang 20222 Bonde 6141 Simbang 2541 Adolang 3180 Sendana Mosso 9211 32773 Putta'da 4619 Sendana 5763 Poniang Tammero'do 7422 Pelattoang Uegamo Onang 3191 Sumakuyu Parabaya Tubo 2567 Baturoro Tubo Malunda Malunda 3671 Malunda 17496 Lombong 4102 Aholeang Kabiraan 2946 Kabiraan Bambangan 2952 U1umandak 3825 Taukong Urekang Pupenga Sources: Kabupaten Po1ewali Mamasa Dalam Angka 1983. Kantor Statistik Kab. Polmas. Majene Dalam Angka 1983. Kantor Statistik Kab. Majene. MANDAR SURVEY page 71 Hap 1: Eabapaten Pol_Ii ••••••a, Soathwlntera SeetiOll, and hbupaten lfajene ••••__ Xabullaten boundery -.-.-K.c .•••t.n bo.und.ry •••.•••••••Da.a/Keluraha" boundary, -Car road ---- root/hor •• trail ~ Ad~lnistrativ. tow~ of Kabupaten CJ • • /"iUag. of Kec•• atan o • village of O•• a/Kelurahan • Other villega. ,.--..., Rivar ~ Distance batw~an vil~ag.s/towns in k•• Sanda"a Na.a o~ .dMinistratlv~',unit • ~ Nallla of .dMinistrativa unit and' Village 72 MANDAR Mandar area corresponds to present day Polewali Mamas a , Majene, and Mamuju kabupatens. (Rahman 1984:40-41, Sahur 1976:16-17,20-23, Kallo 1983:6-7, Isham 1983:16-19) Several suggestions have been made as to the meaning of the word Mandar. As this word in several places is used as a synonym for "binanga" (m river), many claim that it means "river" lt or "water • As one of the main rivers running through the area bears this name, it would not be surprising that it came to be used for the area as well. (Sahur 1976:17) Many historians, however, claim that Mandar is derived from the word Itsipamandaq" , meaning "mutually strengthen", a term used to describe the intention of the Luyo agreement between the Pitu Babana Binanga and pitu Ulunna Salu federations. If this is true, the meaning of Mandar would be "strength". (Sahur 1983:17, Rahman 1984:40) In line with the past unity of the Mandar area, when the Dutch entered the area in 1905, they formed an administrative unit called "Afdeling Manda-ru, which again was divided up into the subdivisions ("Onderafdeling") Polewali, Mamas a, Majene, and Mamuju. The kingdoms were not abolished by the Dutch it was actually not until 1960 that the last kingdom, Balanipa, was fully dissolved but their sovereignty and power were severely curtailed. (Rahman 1984:53,60-62, Kallo 1983:8) After independence a kabupaten Mandar was first established tl in 1952, corresponding to the previous "Afdeling Mandar • Then, in 1959, this was divided into the three present kabupatens: Polewali Mamasa, Majene, and Mamuju. In 1961 the area of the former Balanipa kingdom was divided into the Tinambung (including today's Tutallu), Campalagian, and Wonomulyo kecamatans, while the former Banggae kingdom became kecamatan Banggae. (Kallo 1983:8-9) 1.3. Geography Geographically the area 1S characterized by a coastal plain of varying breadth and a hilly or mountainous interior. ,In kecamatan Wonomulyo the coastal plain extends all the way up to' desa Tapango and in kecamatan Campalagian to the northern part of desa Batupanga. The western part of Campalagian and the eastern part of kecamatan Tinambung is somewhat hilly, however. North of Tinambung, the hills start around the border with Tutallu, getting larger and more rugged toward the interior. Mapilli and Mandar are the two major rivers in this western section of kabupaten Polewali Mamas a , the Mapilli having its source north in the PUS area and the Mandar north in desa Ulumandak in Majene. In Majene the coastal plain is quite narrow, in most places less than a kilometer, and the hills rise immediately to considerable height. Malunda is the only significant river in the kabupaten, also extending into kabupaten Mamuju. MANDAR SURVEY page 73 The coastal plain is basically used for agriculture. The first hills off the coastal plain are rather barren and deforested, but vegetation increases bit by bit toward the interior. In desa Tubbi, for instance, there is rather thick rain forest. 1.4. EcODOmy, Livelihood Farming and fishing occupy most people of the area, although trade and small industry are also of some importance in the main centres. Wet field rice cultivation, a lot of it with artificial irrigation, is dominant in the large coastal plain area of Polewali Mamasa. Along the coast of Wonomu1yo fish farming in ponds also constitutes an important part of the economy. In the other coastal kecamatans of Polewali Mamas a and in all of kabupaten Majene sea fishing plays a more important role. In Majene, due to the rather narrow coastal plain, wet field rice cultivation is not as dominant as in southern Po1ewa1i Mamasa. Although wet field cultivation still accounts for the largest production, dry field cultivation occupies about the same land area. Copra is also an important product in the coastal areas, both inPolewali Mamasa and Majene.· In the interior higher area of kabupaten Majene and kecamatan Tutallu in Polewali Mamasa, the main agricultural products are dry field rice, coffee, and cocoa. Trading of coffee, cocoa, rattan, and other wood products constitutes the main source of cash income in those areas, while the coastal people obtain most of their income from trading of rice, fish, and copra. 1.5. Religion Islam entered the Mandar area in the 16th century. It was first accepted by the Balanipa king and from there spread quickly to the other kingdoms along the coast (Rahman 1984:22). Also today Islam is strongly dominant along the coast as well as in more interior areas such as kecamatan Tutallu and desa Ulumandak. For the area covered by our survey (kabupaten Majene and Tinambung, Tutallu, Campalagian, and Wonomulyo kecamatans of kabupaten Polewali Mamasa) 1983 statistics show 99.7% of the total population to be adherehts of Islam (Majene Dalam Angka 1983, Kabupaten Polewali Mamasa Dalam Angka 1983). There are a few churches in Wonomulyo and one in the town of Majene, but their members are reportedly all immigrants from other areas. 1.6. Education The primary school system seems well developed throughout the area surveyed. Each village visited, even the more remote ones in kecamatan Tutallu, had its own primary school. 74 MANDAR As far as secondary schools are concerned, there is at least one in each kecamatan, usually located at the main centre. High schools of various types (Sekolah Lanjutan Atas SLTA) are located in Majene (7), Somba (1), Tinambung (2), Campalagian (1), and Wonomulyo (2) (Majene Dalam Angka 1983, kabupaten Polewali Mamasa Dalam Angka 1983). The SLTA schools in Polewali also have several students from the area of our survey. The educational opportunities in the area, therefore, must be considered q~ite good. 1.7. Health Facilities General hospitals are found in the towns of Polewali and Majene and several health centres/policlinics in each kecamatan. Where there are roads, mobile health clinics are used as well. People in the coastal area, where transportation is quite easy, must be said to have good access to medical help. In the interior areas, however, the situation is more difficult. In kecamatan Tutallu, for instance, the only health centres are found in Petoosang and Besoangin, which means that it takes up to a day's walk to get medical tratment for many people in that area. With one health centre in the large desa Ulumandak, the situation is not much better there. 1.8. Cam.unication The main road along the coast is paved and in good shape through both kabupatens. Dakka, Batupanga, and Petoosang can also be easily reached by car (preferably 4-wheel drive), although the standard of these back roads is rather poor. From Campalagian it is possible to go about 4 km towards Buku by 4-wheel drive and about 7 km from Wonomulyo. Buku can easily be reached by motorcycle, altho~gh driving from Wonomulyo involves crossing the Mapilli river by raft. The path northwards from Petoosang is not good for anything but foot and horse. No bridges are to be found at any river crossings, so even walking or riding might be difficult in the rainy season. Northwards from Batupanga the path is much better. If carried across a few rivers, it might actually be possible to ride a motorcycle as -far as Rattekallang in the dry season. But walking or riding is probably to be preferred here as well. In kecamatan Tutallu the villages in desa Ambopadang and Tubbi, as well as in the northern part of kelurahan Taramanu, are most easily reached from Batupanga, the villages in desa Besoangin from Pelattoang in kecamatan Sendana, and the other villages from Petoosang. All villages in the kecamatan may be reached in one day or less from one of these three places. This communication situation is also reflected 1n the marketing patterns: Ambopadang and Tubbi people usually go to the MANDAR SURVEY page 75 Mapilli market (via Batupanga), Besoangin people to the Pelattoang market, and the others to the Petoosang market. Reportedly a road from Batupanga to Padangmawalle in the centre of kelurahan Taramanu will be built in the near future. It was also said that when this road is finished, the administrative centre of the kecamatan will be moved from Petoosang to Padangmawalle. If this happens, the communication patterns should change considerably, and it will mean an improvement in communications for those living in the central and northern part of kecamatan Tutallu. As for language, it will probably mean an even greater influence of the Mandar language in these inland areas. In kabupaten Majene the interior villages can only be reached by foot or horse. (Possibly some villages in the interior of kecamatan Banggae may be reached by motorcycle as well.) The northern part of desa U1umandak can probably also be reached in one day. Public transportation is readily available along the ma~n road in both kabupatens both by bus, minibus, and pickup. There are also bus services to Ujung Pandang several times daily. Pickups are available from the main road to Petoosang and Batupanga, although horse and cart seem to be more common on those stretches. 2. SURvEY ITIRBJlAB.Y The survey team, consisting of field linguists Francis.B. Dawson, Kari Valkama, and Kare Str~mme, left Ujung Pandang on motorcycles July 7, 1985, headed for Polewali. The next morning we reported to the Polewali Mamasa Bupati's (kabupaten government head) office (obtaining letters of introduction for the various kecamatans to be visited), as well as to the kabupaten level Depdikbud (Department· of Education and Culture) office and the Kapo1res (kabupaten chief of police) office. We then drove on to Wonomulyo, where we obtained letters of introduction for Palitakan and Buku desas at the Camat's (kecamatan head) office. Having collected wordlists and sociolinguistic information in the villages of Dakka and Buku (both accessible by motorcycle), we went on to Petoosang in kecamatan Tutallu the next day, stopping at the Campalagian Camat's office on the way to ask for a letter of introduction for desa Batupanga. We checked in with the local police and the Camat's office in Petoosang, parked our motorcycles there (end of the road), and setoff for the interior of kecamatan Tutallu on foot aaccompanied by pack horses and a guide. During the following five days we covered the kelurahan/desas of Taramanu, Tubbi, Ambopadang, Pao-Pao, and Allu, collecting word lists and sociolinguistic information in the villages. of 76 MANDAR Lombang, Taramanu, Rattekallang, Tubbi (the Pirian wordlist was also obtained here), Ambopadang, Pao-Pao, and Petoosang. Having finished our work in kecamatan Tutallu, we drove on to the town of Majene the afternoon of July 14. The next morning we reported to Kapolres Majene, the kabupaten Depdikbud office, and the Bupati's office, the latter supplying us with letters of introduction for kecamatan Sendana and Malunda. We then went on to Somba, the administrative village of kecamatan Sendana, where we met with the Camat, obtaining letters of recommendation for the various desas. We also reported to the kecamatan Depdikbud office, where we got helpful information about the language situation in the kecamatan, and to the local police. During this and the following two days we covered the northern half of kecamatan Sendana, collecting wordlists in the villages of Poniang (desa Sendana), Pe1attoang (desa Tammero'do) , Sumakuyu and Parabaya (deBa Onang), and Baturoro and Tubo (desa Tubo). The Rattepadang (desa Tammero'do) and Besoangin (desa Besoangin, kecamatan Tuta11u) word1ists were also obtained in Pe1attoang and the Kabiraan word1ist in Parabaya. On July 17 we continued on to kecamatan Ma1unda, where we also checked in with the local police and the Camat's office. The staff at the Camat's office provided helpful information about the 'language situation and gave us the necessary letters of introduction. During the rest of this and the following day we obtained wordlists from the villages of Aholeang (desa Lombong, collected at Mekatta), Malunda, Kabiraan/Taukong (desas Kabiraan and Ulumandak, - our informant, which we met in the village of Sulae, claimed to represent both villages and said that the speech form was exactly the same in both), Urekang and Pupenga (desa Ulumandak, collected in the village of Malunda). Having finished our work in kabupaten Majene, we headed back south, checking out at the various police offices. A Tinambung wordlist was obtained during a short stop there before going on to desa Batupanga in kecamatan Campa1agian, where we obtained a wordlist from the village of Batupanga at Lena. We also managed to get a wordlist from the village of Lenggo (desa Bu10). Having checked out with the police in Campalagian, Wonomu1yo,and Po1ewa1i, we returned to Ujung Pandang on July 19. 3. HETIIODOLOGY 3.1. Lexical ComparisOD of Wordlists 3.1.1. General To determine tentative language boundaries, we used the method of lexical comparison of wordlists collected in the area. Our definition of language implies mutual intelligibility. So, when we speak of two separate languages, we imply lack of intelligibility. Even though lexicostatistics is not a sufficient MANDAR SURVEY page 77 means of predicting intelligibility, we assume that the measure of lexical similarity roughly correlates with the degree of intelligibility. We feel that lexicostatistics can be used to indicate language relationships, but that any resultant classification must be considered tentative (Smith 1982). Dyen notes: "The reason that lexicostatistical classification is inconclusive would be the same that affects any elassification; not all the knowable facts are known at the time of classification unless one chooses to wait hopelessly until all knowable facts are known. Based on fewer faets than that, a classifieation remains open to correction as additional facts become available" (Dyen 1966:35). The additional facts that make the language classification final would be intelligibility testing between all those language communities which show lexical similarity between 60 to 95 percent. This is because whereas lexical similarity is one eomponent of linguistic similarity (together with phonological, grammatical and semantic factors), intelligibility scores represent linguistic similarity directly, being composite measures of the several factors which determine linguistic similarity. (Simons 1979:15,57,67-87) 3.1.2. Eliciting Data During the survey trip (7-19 July 1985) 27 wordlists were elicited. The earlier Kabiraan wordlist was not used in the calculations because later during the survey we elicited a Taukong/Kabiraan wordlist, which we considered to be more reliable. So, the number of word1ists used in the calculations is 26. We also included seven wordlists from Grimes and Grimes' Languaies of South Sulawesi (to appear), five wordlists from Laskowske and Str~mme's survey of the west-eentral section of kabupaten Polewali Mamasa, and two word lists from Friberg and Kim's survey of the northern part of kecamatan Mambi in kabupaten Polewali Mamasa. These are marked (G), (L/S), and (F/K), respectively, in our matrix. This makes the total number of wordlists 40. The wordlist we used was a 194-item list, which was quite similar to both the Grimes and Grimes and the Laskowske and Str~mme lists, but not identical with either. Out of the 194 items, 192 were also on the Grimes and Grimes list and 191 were on the Laskowske and Str~mme list. All these lists are substantially similar to the Swadesh 200 list. Before eliciting a wordlist, the respondent was screened to ensure that he was a native speaker and the offspring of native speakers of the language/dialect in question. We permitted other local people to advise and correct the respondent, but we required that the respondent himself pronounce the words, so that uniformity of pronounciation could be maintained for a given 78 MANDAR location. Our respondents' age ranged from 23 to 99: 9 were between 20 and 30, 7 were between 30 and 40, 9 were between 40 and 50, 1 was between 50 and 60 and 2 were over 60. 10 of the respondents were farmers and the rest were government employees or religious leaders, e.g. teachers, imams, desa heads and secretaries. Seventeen of our wordlists were elicited in the respective speech community and 9 of them outside of the speech community. The language of elicitation was Indonesian. In two cases though, we had to use interpreters. All our wordlists were elicited by field linguist Kari Valkama. 3.1.3. Lexical Similarity All 40 wordlists were later transferred to a wordbook where each page contained one item in all its 40 varieties. Where a given wordlist elicited more than one response, each response was entered on the wordbook page. After that each word was assigned to its appropriate lexical similarity set according to the principles explained in Bugenhagen (1981:12-14). A lexical similarity set is a set of forms which are all lexically similar to one another (Sanders 1977:36). It is to be noted that here we use the terms lexically similar and lexical similarity set instead of the terms cognate and cognate set, because we want to make it clear that we use the synchronic method instead of the diachronic method. Bugenhagen refers to McElhanon (1967:8), according to whom two items are lexically similar if there is correspondence between fifty percent or more of the phonemes, either as a regular correspondence (for the comparative method) or by phonetic similarity (for the inspection method). (For the discussion of comparative method and inspection method, see Sanders (1977:33-4).) Bugenhagen combines the comparative method with the inspection method to the extent that regular correspondences are taken into consideration. Loanwords, however, he leaves in, which are left out in the comparative method. So, basically we followed the inspection method, with the addition that also regular correspondences were taken into consideration. The decisions were made on a lexically similar/lexically different basis according to the 50% correspondence guideline. 3.1.4. Comparisons and Decisions We agree with Bugenhagen, when he says that the determination of lexical similarity on the basis of phonetic similarity is a somewhat subjective task (Bugenhagen 1981:14). Therefore it seems appropriate to us to give some examples of some of the decisions we made. To start with the easy decisions: 1/2"pohon", elicited the words ponna, poanna, poag, and sappoau, which were all assigned to the same set. Sometimes the chaining of the words made the assignment to a set difficult. For example 41163 "minum", MANDAR SURVEY page 79 elicited the words minum, meiru, meru~, meru, menuQ, menu, minu~, and muitu~. minum and muitu? do not seem to be lexically similar if they stand alone, but with the other words, it is clear that they should be assigned to the same set. 1133 "ayah", elicited the words ama, ambe, ~mbe~, and kama~. ~mbe? and kama? are even farther apart ~n minum and muiru, but ambe is a bridge between the two words, so they were assigned to the-5ame set. 144 "ikan", elicited the words bale and bau, which were assigned to different sets even though ba rs-Iound in both of the words. These few examples show th;t we did not follow the 50% limit mechanically. In the comparisons we compared word roots, not the whole words. So for example the words itunu, mattunu, tunu, mettunu and ditunu were regarded as identical on the base tu~ 3.1.5. Correspondences We found several correspondences between phonemes: #4 h: r:wfJ 19, 170, #101 N+homorganic plosive:long plosive, e.g. nt:tt 114 d:t:r:r 142 s: tl 0 #48 k:t} ISO p:s #20 l:s 157 m: 1 In8 fS:nd 1128 t i s 1138, #171 d:dr r :r : #25 "': o h 3.1.6. Disqualifications Out of the 194-item wordlist, 13 items were disqualified because of ambiguity, lexical repetition, difficulties in lexical similarity set assignment, confusion, and lack of correspondence between the language of elicitation and the language elicited. The items were tHO "sirih", tf24 "mata air", #26 "sungai", 127 "danau", 4,35 "awan", 139 "abu" , 159 "hijau", #88 "tidak", IF98 "timur" , #99 "barat", #129 "laki-laki", 1130 "perempuan", and 1143 "tamu". Eleven additional items were disqualified from some word lists (1-3 items from anyone wordlist) because we strongly suspect that the wrong item was elicited. So, after the disqualifications, the wordlist contained 181 items, and some individual lists even less. 3.1.7. Matrix A matrix containing the 780 comparisons between the 40 wordlists was then produced with the help of a computer. (This matrix is not included here.) Because each lexical similarity percentage indicates a range rather than a specific value, the 80 MANDAR ranges of two different lexical similarity percentages may overlap. Therefore the next step was to reduce the matrix to its significant differences. In this process we followed Gary Simons (1977:75-105). We used the table for 200 words and confidence level .10, which is used in average survey situations with good bilingual respondents, as was the case in this Mandar survey. In the reduction percentages between: 52-54 were reduced to 53 55-61" " " 58 62-68 tI •• " 65 69-74 II " tI 72 75-80" •• " 77 81-85 II •• •• 83 86-90" " " 88 91-94 II " It 92 96-98 II " n 97 There were no percentages below 52 or above 98, neither did percentage 95 occur. The resultant matrix is shown in Table 2. 3.2. sociolinguistic Questionnaire In addition to our main method of collecting wordlists for lexical comparison, a sociolinguistic questionnaire in Indonesian was used in 10 villages to obtain additional information. The questionnaire included questions on communications and patterns of movement, economy, trade, religious and educational situation, health facilities, social interaction, as well as on language use patterns and language differences. The questions on language differences, where the informant was asked to group the villages of the area according to their speech form on a scale from precisely the same as his own to totally different, were particularly helpful as a guide on where to collect wordlists and on where exactly to draw the language boundaries. The questions on language use patterns gave us important information on the vitality of the local languages. The other information obtained has already been incorporated under section 1 above. Most of our informants for the sociolinguistic questionnaire were natives of the villages concerned and all had extensive knowledge of the local situation. Most of them were village or desa leaders or had other prominent positions in the society. In some cases they were assisted or corrected by other villagers present. Information obtained by direct questioning does not always correspond to reality. Nevertheless, we have found the result of our lexicostatistic comparison to largely confirm the reported language variation, and to the extent that we were able to observe language use, those reported patterns were also confirmed. MANDAR SURVEY page 81 Table 2: a.e.tucedHatrix of Lexical Siailarity Perceatqea and LaDgaage OroupiDc- 82 4. RESULTS 4.1. Language Boundaries According to Grimes and Grimes' method of classification, which we will also follow, lexical similarity percentages above 80 are considered to indicate one language. Percentages between 75 and 80 are considered to indicate the same subfamily of languages. Percentages between 60 and 75 are considered to indicate the same family of languages. As already noted, lexical similarity percentages are not sufficient to differentiate between dialects, so we did not attempt a breakdown of languages into dialects. Based on the figures mentioned above, we arrived at a linguistic configuration of the area as described in Table 2 and Map 2. Five languages, with boundaries approximately as indicated on Map 2, are spoken in the frea covered by our survey: Dakka, Campalagian, Mandar, Ulumandak, and PUS (represented by our Pup~ga wordlist). The other languages included in our matrix and map, pan~ei and Aralle-Tabulahan (if this is really a separate language), represent areas covered by Laskowske and Str~mme's 1984 survey. PUS, Aralle-Tabulahan, Ulumandak, and Pannei all relate to each other at an average of 75-80% and have therefore been grouped into one subfamily named the PUS Subfamily in accordance with Str6mme's report of the survey mentioned above. Two language families are represented in our matrix: the Northern South Sulawesi Family (represented by the PUS Subfamily and the languages of Dakka and Mandar) and"the Bugis Family (represented by the Campalagian language). A few comments are needed on the grouping of some of the Ulumandak wordlists. Both Tubbi and Kondo relate to both the Ulumandak and Pannei lists at an average of above 80%, but as they show higher average lexical similarity with the Ulumandak lists, we have chosen to include them with the Ulumandak language. Despite previous researchers' grouping of Tapalang with Mamuju, we have grouped Orobatu (Grimes and Grimes' Tappalang list) with the Ulumandak language, as its average lexical similarity with the Ulumandak lists is 81.4%. According to Grimes and Grimes' matrix, this list shows an average lexical similarity of 82.2% with the Mamuju lists. Grimes and Grimes have therefore included Tappalang with the Mamuju language. A new and more complete matrix of the Mamuju language, however, made by field linguists T. Laskowske and K. Valkama after a recent survey in kabupaten Mamuju, shows Orobatu to relate at only 80.2% with the Mamuju lists. Orobatu obviously represents a central link between the two groups, but as the latter matrix is more complete for the Mamuju language than that of Grimes and Grimes and as the principles applied by Laskowske and Valkama in the assignment of MANDAR SURVEY page 83 IIap 2: LaD.aua~'. of Eab1apaten Polt!lnlli "-··at S_thwaatera Sectioa , mul EabapateD Kaje:ae ra••Uy boundary SubraMily boundary Language boundary Possible· languaga .boundary. Kabupatan boundary Villaaa/town rapra.entad with wordliat in lIlatrio< Othar town/village Languaga nama "i ••d language area with ••any i•• igrenti . Although w'·~.va choaen to include n Kab. "a"uJu with tha Ulu ••andak OO>AR drawing the boundary batwaan tha and tha ~alllujulanguaga i. bayond 84 •••,. •.•h.t u •••••••• u.P ••••••••••. f A ••••••.••••• lexical similarity were the same as those applied for this report, we have chosen to group Orobatu with Ulumandak. As far as language boundaries are concerned, all of the earlier stated aims for this survey have been reached: The northern boundary of the Mandar language 1n kabupaten Polewali Mamasa has been established, a language in the area between Mandar and PUS, Ulumandak, has been found to exist and its boundaries in kabupaten Polewali Mamasa and Majene approximately established. Dakka has been found to be a separate language and Buku to be part of the Campalagian language. As to Sendana, the wordlists from this kecamatan do not relate to each other closer than 88% at an average. It seems more likely that there are several dialects within this kecamatan. It is certainly not right to speak of a Sendana dialect extending into kecamatan Malunda. Awo' Sumakuyu has been found not to be a dialect of Mandar, but part of the Ulumandak language. The Malunda dialect was reported to be spoken only by part of the people in the village of Malunda and its immediate surroundings. Except for a number of Mandar speaking bmnigrants from areas further south who have settled in the villages llmmediatelynorth of the village of Malunda, kecamatan Malunda consists of Ulumandak speakers. We have been able to verify that the Ulumandak language extends to the border of kabupaten Mamuju, including the coastal villages of Mekatta and Maliaya. It was also reported to extend across the border, but how far we were not able to check on this survey. Thus, the Mandar and Mamuju languages do not have a common boundary but are separated by an Ulumandak speaking area. (Table 3 summarizes the languages in Polewali Mamasa and Majene and gives the estimated number of speakers for each language.) 4.2. Language Use Patterns Except for Ulumandak, we did not use any formal questioning about language use in kabupaten Majene. The vitality and dominant position of the Mandar language in that area seemed all too obvious to make any such questioning necessary. For the areas where we did use the questionnaire (desa Ulumandak, kecamatan Tutallu, and the villages of Dakka and Buku) the reported situation may be summarized as follows: Apart from a few exceptions the local language is used exclusively in the home. Some Indonesian was reported to be used in the home in the villages of Pao-Pao, Ambopadang, and Dakka, in the latter case only when children were involved. At several places some Indonesian was reported to be used by school children when playing. MANDAR SURVEY page 85 Table 3: LaDgaagea of Kabapaten Polewali x....a, Southwestern Sec:ti01l, &IUl hbapaten Jlajeae with Eatu.ated Ru.ber of Speakers Estimated Number Language of Speakers Pannei 9000 Dakka 1500 Campalagian 30000 Mandar 200000 Ulumandak 18000* pitu Ulunna Salu 300** The estimates are based on ou~ findings as to the extent of the language, sociolinguistic information obtained, and the population figures in table 1. * Figure does not include U1umandak speakers in kab. Mamuju. ** Figure only includes the village of Pupenga in Ulumandak. Total number of P.U.S. speakers is estimated to be 22000. 86 MANDAR In situations like working together with fellow villagers in the rice fields, the local language was reported to be used exclusively in every village apart from Ambopadang, where some Indonesian was also claimed to be used. As for the language use at markets, government offices, and health centres, the answers were more varied. It seemed to vary both· with the location of these facilities and with the people dealt with at these places. With a few exceptions Indonesian and the Mandar language were claimed to be used throughout. Speakers of Ulumandak and Campalagian (at Buku) also claimed to use the local language when these facilities were located within the local area and when dealing with people of their own group. Dakka stands alone in reporting only Indonesian to be used. This may be explained by the fact that this is a very mixed language area. In fact only 20% (about 1000 speakers) of the inhabitants of desa Palitakan, where most Dakka speakers live, were reported to speak the Dakka language, the rest being a mixture of mainly Bugis and Mandar speakers. As for the religious context, a mixture of Arabic, Indonesian, and the local language is used in the mosques. Arabic only is used for praying. The preaching is most often in Indonesian, but the local languages are also used to some extent. Announcements in the mosques are most often given both in Indonesian and the local language. In Pao-Pao they were actually reported to be given in the Mandar language only. For more culturally related activities such as traditional feasts (where these still exist) and ancestral storytelling, the rule is that the local language is used. At Ambopadang Indonesian was also reported to be used for storytelling due to a limited understanding of the local language among children. In the Ulumandak-speaking village of Tubbi the Mandar language was also said to be used to some extent at traditional feasts due to some Mandar speaking immigrants from the south. When it came to such an intimate, personal thing as expressing anger at other people, the answers were unanimous: Only the mother tongue is adequate for that. We did not have the opportunity to properly check these reported language use patterns, but at least they did not seem to disagree with our limited and casual observations. Despite a fairly widespread knowledge of Indonesian, especially in the coastal areas, it seems safe to say, at least as a preliminary conclusion, that overall the languages of Mandar, Campalagian, and Ulumandak are vital and have a strongly dominant position in their respective areas. Neither are there any signs of change in that situation. As the major language of the region, Mandar also seems to exert some influence in the Campalagian, Ulumandak, and Dakka speaking areas, though, as a regional trade language. Throughout. MANDAR SURVEY page 87 the area covered by our survey, Mandar vas claimed to be well understood, despite the lack of mutual intelligibility. As to Dakka, the situation is more unclear. As already noted, there is a strong aixture of ethnic backgrounds and languages represented in that area, with the result that Indonesian more and more seems to be taking over as the language of ordinary communication. The children of Dakka parents were already reported to have a limited understanding of the Dakka language, a trend that the parents did not seem too eager to oppose. But a final judgement about the future of this language should not be made before a more thorough sociolinguistic investigation has been undertaken. 4.3. ca.pariaoa with Prerioua Work As nothing has been written before on language use patterns, the following comparison only deals with language divisions and boundaries. Many similarities can be pointed out between the results of our survey and previous descriptions of the area, as presented under section 0.1., but also a number of disagreements. We agree with all previous researchers that Mandar is a separate language from PUS. Our boundary between the Mandar language and what we have called the PUS Subfamily is also not too different from that between the Mandar and PUS (or Sa1dan) languages (or subgroups) of previous works. East of the Mapilli river our Mandar boundary follows very closely that of van der Veen and Peta Bahasa, following a course south of Dakka. We disagree with them, however, that Dakka should be grouped with pus. Although Dakka's closest relationship is with the Pannei language of the PUS Subfamily (72-77%), we have found its lexical similarity with other members of that subfamily too low to include it. At the Mapilli river our Mandar boundary takes a sharper turn north than suggested by previous works, approximately following the course of the river all the way up to the village of Rattekal1ang, including this and some villages on the east side of the river, like Lenggo, within the Mandar language. From Rattekallang it makes a turn to the south, including the village of Taramanu with the PUS Subfamily, before continuing west to the Majene coast at Poniang. Contrary to van der Veen, then, we have found all of desa Ambopadang, most of ke1urahan Taramanu (everything south of Taramanu village), and part of desa Tubbi (the village of Ratteka1lang and those south of it) to belong to the Mandar language. Our boundary here comes closest to that of Peta Bahasa and Grimes and Grimes, but ours is still somewhat further to the north. 88 MANDAR In the west we have found that the Mandar language does not extend as far inland as suggested by previous researchers. It generally seems to be limited to the narrow coastal plain, while the hills above are inhabited by Ulumandak speakers, some of whom have even moved down to the coast and settled in the villages of Poniang, Sumakuyu, Tubo, Sulae, and Mekatta. Our inclusion of the Ulumandak area within the PUS Subfamily corresponds with previous researchers' grouping of this area. Our suggestion of an Ulumandak language as separate from the PUS language and comprising a much larger area than desa Ulumandak, confirms Str~mme's anticipations as expressed in his 1985 report, but the exact extent of this language in kabupatens Polewa1i Mamasa and Majene has not been established until now. Similar to Peta Bahasa but contrary to other previous works, we have found the Mandar and Mamuju languages not to have a common boundary. Peta Bahasa describes Botteng-Tappalang (under which the Orobatu wordlist in our matrix belongs) as a separate subgroup (apparently corresponding to our use of the term language) with origins in the PUS area. This description seems to fit quite well with our inclusion of Orobatu with the Ulumandak language and the PUS Subfamily. We also note that our northern boundary of the Mandar language is very similar to that of Grimes and Grimes. Due to a very mixed population in that area, the eastern boundary of the Mandar language in kabupaten Po1ewali Mamasa seems very difficult to determine. We have drawn it close to the border between kecamatan Wonomulyo and Po1ewali, which seems to be what is preferred by Peta Bahasa. Possibly it extended further east before the immigration to the area started. Now it might be more accurate to draw the boundary west of the immigration area and say that some Mandar speakers are also found among the immigrants to the east. The precise extent of the Pattae' language and the southern boundary of the Mamas a language is also somewhat uncertain. To get the full understanding of the linguistic situation in this southeastern section of kabupaten Polewa1i Mamasa another survey is needed. Although several previous researchers have described Pattae' (Binuang) as a dialect of Mandar, we do not find it necessary to do any further checking on that. Grimes and Grimes' lexical similarity percentages show quite clearly that this is a separate language. With regard to Campalagian, our survey has confirmed Grimes and Grimes and Peta Bahasa's disqualification of it as a dialect of Mandar. We have also confirmed Peta Bahasa's inclusion of the village of Buku with this language. As to the classification of Campalagian, we have followed Grimes and Grimes in describing it as a separate language in the Bugis Family rather than a dialect of it, as suggested by Peta Bahasa. As we have mentioned earlier, it was not our intention to do a dialect survey of the Mandar language. Nevertheless, our MANDAR SURVEY page 89 findings make a reV1S10n of the earlier statements about the dialect situation of Mandar necessary, as follows. As seen from our matrix, our Sumakuyu wordlist fits easily into the Ulumandak language, and the village of Awo was said to have the same speech form. We can therefore no longer support Peta Bahasa and Grimes and Grhnes' suggestion of Awo' Sumakuyu as a separate dialect of Mandar. If it is a separate dialect, it is a dialect of the Ulumandak language. Assuming that it must be the villages of Awo and Sumakuyu they have in mind when they talk about the Awo' Sumakuyu dialect,' the locations given by Peta Bahasa and Gr~es and Gr~es for this dialect also seem a bit off. The former says it is located in kecamatan Ma1unda and the latter on "the peninsula protruding west above the Sendana area", i.e. in the desas of Tubo and the northern part of Onang in kecamatan Sendana. The village of Awo, however, we have found to be located in the northern part of desa Tammero'do and Sumakuyu in the central part of desa Onang, both in kecamatan Sendana. (See Map 1.) We therefore suspect that neither Grimes and Grimes nor the editors of Peta Bahasa knew exactly where these villages were located. With regard to the other dialects of Mandar, it looks as if we have at least confirmed Grimes and Grimes' classification of Malunda as a separate dialect. As the two Ma1unda lists relate to the other Mandar lists as low as 76-85%, it seems safe to say this despite our reluctance to distinguish dialects on the basis of lexicostatistics. Being limited to the village of Malunda and its immediate surroundings, the Malunda dialect does not cover an area as large as indicated on Grimes and Grimes' map, however. Regarding Sendana we have found, contrary to previous works, that this dialect does not extend beyond kecamatan Sendana to the north. If Sendana speakers are also found in kecamatan Malunda, these are probably people who have moved up from the south. As has already been mentioned, our findings also raise the question whether it is right to speak of only one Sendana dialect as all previous works do. Our lexical similarity figures for the wordlists taken in the Mandar speaking area of kecamatan Tinambung, Tutallu and Campalagian also raise the question whether it is right to speak of only one dialect, Balanipa, for the whole of that area as all previous works do. On the other hand one may ask whether Grimes and Grimes are right in suggesting two different dialects, Ba1anipa and Majene, on the basis of two wordlists relating to each other at 97% according to our matrix. Could it be that what has previously been called the Balanipa dialect is actually located to the north and/or east of the villages of Napo and Tinambung? It is interesting to note here, by the way, that with regard to the wordlist on which the first suggestion of a Ba1anipa dialect is based, that of Adriani and Kruijt, Mills concludes that 90 MANDAR it represents "either a dialect transitional between Mandar and PUS or, more likely, simply a dialect of PUS which happened to be spoken within the (then) administrative unit called Ba1anipa. This would be the northernmost inland part of the area." Mills own Ba1anipa list, though, which originates from the coast, he has found to share very similar phonological traits with Adriani and Kruijt and Pe1enkahu's Majene data, traits that we find to be even more similar when comparing our and Grimes and Grimes' data from these places. S. RESIDUES - PoR'tREk. RESEARCH REEDED In relation to the stated aUns for this survey, no further residues can be said to exist. That does not mean, however, that there are no further residues with 'regard to the linguistic configuration of the area surveyed. In fact, our survey has not brought us much further in meeting what Mills calls "the most pressing need" with regard to the Mandar language, i.e. "a thorough dialect survey to determine as far as possible the exact boundaries and isoglosses." Our findings in kecamatans Sendana, Tutallu, and northern Campalagian have raised the question whether it is right to speak of only one Sendana and only one Balanipa dialect. It also needs to be demonstrated why Ba1anipa and Majene are separate dialects, if that really is the case. Our survey has also showed that the need for such a dialect survey is not any smaller when it comes to the U1umandak language. Our figures here could well indicate that each word list in our comparison represent a separate dialect. Even within the small area of the Campalagian language our Buku word list would seem to indicate more than one dialect. A future dialect survey of the area should not only seek to determine the boundaries of the major dialects, but also to describe and locate all the minor variants within each of these, which, according to Peta Bahasa, there are quite a number of. As mentioned before, a thorough dialect survey will require other methods than those applied for this survey. Lexicostatistics is just not sufficient for that task. 6. CORCLUSIOR Six languages are spoken by the nonLumigrant inhabitants of the southwestern section of kabupaten Polewa1i Mamas a and kabupaten Majene: Campalagian, Mandar, Dakka, Pannei, Ulumandak, and pitu U1unna Sa1u (the village of Pupenga only). Of these at MANDAR SURVEY page 91 least Campalagian, Mandar, Pannei, and Ulumandak are represented with more than one dialect. As far as linguistic descriptions of these languages are concerned, a fair amount is already available on the Mandar language and a little on Campalagian. (See bibliography.) Also, an UNHAS-SILfield research program has been started in the PUS language. No studies have as yet been undertaken in the Dakka, Pannei, and Ulumandak languages. According to our analysis of the situation, there seems to be a need for a full UNHAS-SIL field program in at least Campalagian, Pannei, and Ulumandak. Further investigation of sociolinguistic factors is needed to decide the needs with regard to Dakka. As to Mandar, at least the dialect situation requires further research. Whether the available linguistic descriptions of that language and whether the reading materials available in it are sufficient and adequately cover all dialects, still needs to be determined. lfOTES 1. Only people from desas Ulumandak and Kabiraan use this name for their language. At most other places within this group people named their language with the name of their village. We have chosen to use Ulumandak as the name for the whole language, as the Ulumandak area seems to have a central position both linguistically and geographically. Whether this name would be accepted by the speakers throughout the language group, remains to be tested. 2. As Aralle shows an average of 82.3% lexical similarity with the PUS wordlists of Saludadeko, Galung and Pupenga, and as Str~mme previously has concluded that Aralle-Tabulahan and PUS are two separate languages, a few comments are needed concerning the status of Aralle-Tabulahan: 1. We threw out 13 words of the 194-item wordlist mainly because we suspected that they elicited erroneous responses. That raises the percentage counts because suspected items are those which are different from each other. 2. When our lexical similarity percentage counts are compared with those of Str~mme's 1985 Polewali Mamasa report, it appears that for the lists shared by both (i.e. the Saludadeko, Galung, Aralle, Kondo, Karombang, Pulliwa, Tapango, Napo (Balanipa) and Malunda lists), our figures are 3.7% higher on average. 3. When wordlists are taken from speech communities where other surveyors have taken wordlists before, the differences are sometimes surprisingly great. Grimes and Grimes' Malunda wordlist, for example, is only 92% lexically similar with our 92 MANDAR Malunda wordlist. Therefore the percentage figures cannot be fully trusted. 4. Previous surveyors have come to contradictory conclusions regarding the Aral1e-Tabulahan group. Grimes and Grimes describe it as a dialect of PUS, while Str~mme classifies it as a separate language. Friberg and Kim do not seem to make any conclusion as to its classification~ 5. The focus of our survey was not on Ara1le-Tabu1ahan or PUS, located in the northern part of kabupaten Polewali Mamasa, but on the southern part of this kabupaten and on kabupaten Majene. Therefore we have included only one of the five Aralle-Tabulahan wordlists available and only three of the 12 PUS wordlists. Our survey therefore does not give the total picture of the linguistic relationship between PUS and Ara1le-Tabu1ahan. However, as our percentages are on average 3.7% higher than those of Str~e's report and as the percentages of that report for the other Aralle-Tabulahan lists in relation to the PUS lists are on average not lower than between the lists we have included, we can probably not expect the picture to have been more in favour of a separate Ara11e-Tabu1ahan language if all these lists were included in our comparison. Because of the above facts, since generally speaking the method of counting lexical similarity itself gives only a tentative picture of the language situation, and since there is an UNHAS-SIL team assigned to do in-depth research in the PUS language, we want to leave the question open, whether Aral1e-Tabulahan is a dialect of the PUS language or a language of its own. Further research, especially intelligibility testing, will give the final answer. MANDAR SURVEY page 93 MANDAR BIBLIOGBAPHY A listing of linguistic, anthropological, and various other works relating to the Mandar area (kabupaten Polewali Mamas a , Majene, and Mamuju). Adriani, N., and A.C. Kruijt. 1914. De Bare'e-sprekende Toradja's van Midden-Celebes. Derde dee1: Taa1- en 1etterkundige schets der Bare'e taal en overzicht van het taalgebied: Celebes - Zuid-Ha1mahera. Batavia: Landsdrukkerij. A1iah, R.St. n.d. Kerajaan Banggae Majene. Article collection. Majene: Seksi Kebudayaan, Kandep P & K. Anonymous. 1909. Mededeelingen betreffende eenige Mandharsche landschappen. Bijdragen tot de Taa1-, Land- en Vo1kenkunde van Nederlandsch-Indie 62.649-746. ----. 1981. Laporan perjalanan penelitian dan pemetaan bahasa-bahasa daerah di Sulawesi Selatan. Ba1ai Pene1itian Bahasa, Ujung Pandang. MS. ----. 1981/82. Upacara tradisional daerah Sulawesi Selatan. Proyek Invetarisasi dan Dokumentasi Kebudayaan Daerah Sulawesi Selatan. Ujung Pandang: Departemen P & K. ----. n.d. Kabupaten Po1ewali Mamasa dalam angka 1983. Kantor Statistik, Polewali. MS. ----. n.d. Kecamatan Tinambung da1am angka 1983. Mantri Statistik, Tinambung. MS. ----. n.d. Laporan hasil seminar kebudayaan Mandar I tangga1 31 Juli s/d 2 Agustus 1984 di Majene. Panitia Seminar Kebudayaan Mandar I, Ujung Pandang. MS. ----. n.d. Majene dalam angka 1983. Statistik tahunan. Kantor Statistik, Majene. MS. ----. n.d. Mamuju dalam angka 1981/82. Statistik tahunan. Kantor Statistik, Mamuju. MS. Ba'dulu, Abd. Muis. 1980. Interferensi gramatikal Bahasa Mandar dalam Bahasa Indonesia murid sekolah dasar di Sulawesi Selatan. MS. ----. 1979/80. Morfologi dan sintaksis Bahasa Mandar. Research report, Fakultas Keguruan Sastra Seni, IKIP Ujung Pandang. MS. ----. 1985. Sistem morfologi kata kerja Bahasa Mandar. Jakarta: Pusat Pembinaan dan Pengembangan Bahasa, Departemen P & K. Banru, A. Talib. 1976. Todilaling sebagai pembaharu kerajaan Balanipa Mandar. B.A. thesis, Hasanuddin University. Bataragoa, Vivia. 1975. Tinjauan historis tentang masuk dan berkembangnya agama Kristen/Protestan di daerah Mamasa. B.A. thesis, IKIP Ujung Pandang. 94 MANDAR Buntugayang, Samua1. 1976. Satu pengungkapan historis mengenai perjuangan Demmatande me1awan imperia1isme Be1anda tahun 1914-16 di Pitu U1unna Sa1u, kabupaten Po1ewa1i Mamasa. Thesis, IKIP Ujung Pandang. Djama1uddin, Husni. 1984. Kesenian Mandar dan masa1ah kreatifitas. MS. Djubaer, Arfah Adnan. 1974. Tinjauan puisi Mandar (Ka1inda'da) dan sumbangan terhadap puisi Indonesia. Thesis, IKIP Ujung Pandang. Esser, S.J. 1938. Ta1en. In Atlas van Tropisch Nederland. Amsterdam: Konink1ijk Neder1andsch Aardrijkskundig Genootschap. Friberg, Timothy. UNHAS-SIL sociolinguistic survey: Kabupaten Po1ewa1i Mamasa, northern section. In this volume. ----, and Barbara A. Friberg. 1984·. A dialect geography of Bugis. Summer Institute of Linguistics, Indonesia Branch. MS. Graff, S. De, and D.G. Stibbe. 1918 Ceds.) Encyc10paedie van Neder1andsch-Indie. Tweede dee1. Grimes, Charles E., and Barbara D. Grimes. 1984. Languages of South Sulawesi. To appear. Hoorweg. 1911. Nota bevattende eenige gegevens betreffende het landschap Mamoedjoe. Tijdschrift voor Indische Taa1-, Land- en Vo1kenkunde (Tijd. van het Bataviaasch Genootschap van Kunsten en Wetenschappen) LIII.57-154. Isham, Husniah. 1983. Aspek arkeologis makam raja-raja Banggae di Ondongan kabupaten Majene. Thesis, Hasanuddin University. Ka110, Abdul Madjid. 1981. Orang Mandar dan eko1oginya. Suatu studi sistim eko1ogis orang Mandar di desa Labuang kabupaten Majene. Thesis, Hasanuddin University. ----. 1981/82. Model-model unit produksi petani ne1ayan Mandar di sebuah desa pantai Mandar. Research project, Hasanuddin University. MS. ----. 1982. Tekno1ogi petani/ne1ayan Mandar. Satu studi tekno1ogi tradisiona1 di bidang perikanan dan pertanian di desa pantai Labuang, kabupaten Majene. Research project, Hasanuddin University. MS. ----. 1983. Sistem tekno1ogi dan per1engkapan hidup orang Mandar di Sulawesi Se1atan. Hasanuddin University. MS. Kruijt, A.C. 1942. De bewoners van het stromgebeid van de Karama in Midden-Celebes. Neder1andsch Aardrijkskundig Genootschap, 2e Serie dee1 LIX, Af1. 4. Leiden: E.J. Brill. Laskowske, Thomas V. UNHAS-SIL sociolinguistic survey: Seko area. In this volume. Lopa, H. Baharuddin. 1984. Sirik dalam masyarakat Mandar dan pemanfaatnya da1am pembangunan di Sulawesi Se1atan. Pusat Pene1itian Pembangunan Pedesaan dan Kawasan Lembaga Pene1itian, Hasanuddin University. MS. MANDAR SURVEY page 95 Mangemba, H.D., et a1. 1978/79. Sastra lisan Mandar. Research report, Proyek Penelitian Bahasa dan Sastra Indonesia dan Daerah Sulawesi Selatan, Ujung Pandang. MS. Mills, R.F. 1975. Proto-South Sulawesi and proto-Austronesian phonology. Doctoral thesis, University of Michigan. Muthalib, Abdul. 1973. Dialek Tallumpanuae atau Campalagian. Ujung Pandang: Lembaga Bahasa Nasional Cabang III. ----. 1977. Kamus Bahasa Mandar - Indonesia. Jakarta: Pusat Pembinaan dan Pengembangan Bahasa, Departemen P & K. ----. 1985. Perencanaan penyusunan kamus bahasa daerah. Paper presented at Pertemuan Bahasa dan Sastra Daerah Wilayah Timur, Ujung Pandang, January, 1985. MS. ----. 1984. Kedudukan dan fungsi Bahasa Mandar. Proyek Penelitian Bahasa dan Sastra Indonesia dan Daerah Sulawesi Selatan, Ujung Pandang. MS. ----. 1984. Sistem perulangan Bahasa Mandar. Jakarta: Pusat Pembinaan dan Pengembangan Bahasa, Departemen P & K. Pelenkahu, R.A. 1967. Gambaran sepintas la1u tentang dialek-dialek Mandar. B.A. thesis, IKIP Ujung Pandang. ----. 1974 (eds.) Peta bahasa Sulawesi Selatan (Buku petunjuk). Ujung Pandang: Lembaga Bahasa Nasional Cabang III. ----. 1975. Loka karya pembakuan ejaan latin bahasa-bahasa daerah di Sulawesi Selatan (25 sId 27 Augustus 1975 di Ujungpandang). Balai Penelitian Bahasa, Ujung Pandang. MS. ----. 1975. Seminar pembakuan ejaan latin bahasa-bahasa daerah di Sulawesi Se1atan (26 sId 29 Maret 1975 di Ujungpandang). Lembaga Bahasa Nasiona1 Cabang III, Ujung Pandango MS. ----. 1983. Struktur Bahasa Mandar. Jakarta: Pusat Pembinaan dan Pengembangan Bahasa, Departemen P & K. Pua1illin, Dorkas. 19790 Tinjauan historis tentang seni budaya di daerah Kondo Sapata' Mamasa. B.A. thesis, IKIP Ujung Pandango Rahman, Sukirman A. 1984. Sejarah daerah Majene. Ba1ai Kajian Sejarah dan Nilai Tradisiona1, Ujung Pandang. MS. Roisuddin. 19820 Sistem perkawinan adat Mandar di Pambokborang desa Toto1i, kecamatan Banggae, kabupaten Majene. Research report, Pusat Latihan Pene1itian Ilmu-I1mu Sosiat, Hasanuddin University. MS. Sahur, Ahmad. 1976. Penelitian ni1ai-nilai budaya dalam kesusastraan Mandar. Research report, Faculty of Letters, Hasanuddin University. MS. ----0 1979. Perahu Sande sebagai perahu spesipik orang Mandar. Thesis, Hasanuddin University. ----. 1984. Kamus sederhana Bahasa Mandar - Indonesia. Ujung Pandang: Ikatan Keluarga Wanita Po1emaju Mandar. ----, and Majid Kallo. 1982/83. Masalah bendi di Mandar. Thesis, Hasanuddin University. 96 MANDAR . Salim, Amri. 1976. Tinjauan historis tentang pertumbuhan dan perkembangan agama Islam di daerah Majene. B.A. thesis, IKIP Ujung Pandang. Salzner, Richard. 1960. Sprachenatlas des lndopazifischen Raumes. Wiesbaden: Otto Harassowitz. Samar, Abd. Azis, and Arianus Mandadung. 1979. Ungkapan sejarah dan budaya di kabupaten Polewali-Mamasa, Sul-Sel. -Seri: "All s (Daerah Kondosapatar/Mamasa). MS. Sangi, M. Zaino 1972. Tinjauan sintaksis dialek Balanipa Mandar menurut tata bahasa transformasi. Thesis, IKIP Ujung Pandang. Sikki, Muhammad H. 1985. Kata tugas atributif Bahasa Mandar. Paper presented at Pertemuan Bahasa dan Sastra Daerah Wilayah Timur, Ujung Pandang, January, 1985. ----. 1985. Kata tugas Bahasa Mandar. Proyek Penelitian Bahasa dan Sastra Indonesia dan Daerah Sulawesi Selatan, Ujung Pandang • MS • Str~mme, Kare J. UNHAS-SIL sociolinguistic survey: Kabupaten Polewali Mamasa, west-central section. In this volume. Syah, M.T. Azis. 1979. Terjemahan Kalinda'da Mandar. Vol. I-II. MS. ----. 1980. Apresiasi sastra Mandar. MS. ----. 1983. Dasar-dasar sastra Mandar. MS. ----. 1984. Biografi I Calo Ammana I Wewang Topole Dibalitung,. pahlawan daerah Mandar Sulawesi Selatan. ----. 1985. Fungsi sastra daerah Mandar Sulawesi Selatan bagi masyarakat pemakainya. Paper presented at Pertemuan Bahasa dan Sastra Daerah Wilayah Timur, Ujung Pandang, January, 1985. MS. Tangdilintin, L.T. 1984. Ungkapan tradisional yang ada kaitannya dengan sila-sila dalam Pancasila propinsi Sulawesi Selatan. Jakarta: Departemen P & K. Tenriadi, A. 1961. Hikajat tanah Mandar. Bahasa dan Budaya 9:1.18-33. ----, and G.J. Wolhoff. 1955. Lontara Mandar. Bahasa dan Budaya 3:3.7-31, 3:4.25-38. Veen, H. van der. 1929. Nota betreffende de grenzen van de Sa'dansche taalgroep en het aanverwante taalgebied. Tijdschrift voor Indische Taal-, Land- en Volkenkunde 69.58-97. Yasil, Surady. 1978. Selayang pandang beberapa tema puisi Mandar. Thesis, Hasanuddin University. ----. 1982. Kalindaqdaq Mandar dan beberapa temanya. Ujung Pandang: Balai Penelitian Bahasa. Yuseng, Muhammad. 1983. Sumbangan pendidikan Islam terhadap perjuangan mendukung proklamasi 17 Agustus 1945 di Mandar. Thesis, Hasanuddin University. MANDAR SURVEY page 97 :.4 .1:~;~,' .s A listing of other articles referred to in this report. Arden, G.S. 1977. Guidelines for conducting a lexicostatistic survey in Papua New Guinea. In Loving 1977. Bugenhagen, Robert. 1981. A guide for conducting sociolinguistic surveys in Papua New Guinea. Summer Institute of Linguistics. Dyen, Isidore. 1966. Comment. Oceanic Linguistics 5:1.32-49. Loving, Richard. 1977 (ed.) Language variation and survey techniques. Workpapers in Papua New Guinea Languages 21. McElhanon, K. 1967. Preliminary observations on Huon Peninsula languages. Oceanic Linguistics 6.1-45. Simons, Gary F. 1977. Tables of significance for lexicostatistics. In Loving 1977. ----. 1979. Language variation and limits to communication. Technical report no. 3. Ithaca: Cornell University. Smith, Kenneth D. 1982 (ed.) A compendium of articles relating to the survey of Sabah languages. Pacific Linguistics. 98 MANDAR URBAS-SIL sociolinguistic Survey: Kabupaten Haauju Kari Valkama TABLE OF COB"tlUftS page INTRODUCTION 100 1. NONLINGUISTIC INFORMATION 100 2. METHODOLOGY 101 3. RESULTS 105 4. RESIDUE 110 5. BIBLIOGRAPHY III APPENDICES 113 MATRICES: Matrix 1. Central Sulawesi Stock 105 Matrix 2. South Sulawesi Stock 108 Matrix 3. West Austronesian Superstock 113 MAPs: Map 1.a. Kabupaten Mamuju, Northern section 114· Map 1.b. Languages of Kabupaten Mamuju, Northern section 115 Map 2.a. Kabupaten Mamuju, Southern section 116 Map 2.b. Languages of Kabupaten Mamuju, Southern section 117 MAMUJU 99 :nrrRODIJCTIOR The survey was conducted in three parts during the period September 19 - November 14, 1985. The first trip was made by Tom Laskowske and Kari Valkama September 19 - October 2 to kecamatan Kalumpang and from there up the coast to kota kecamatan Pasangkayu. The second trip was made by Don Barr and Kari Valkama October 18 - 25 to kecamatan Pasangkayu. The third trip was made by Kare Str;mme and Kari Valkama November 6 - 14 to kecamatan Budong-Budong, Kalukku, Mamuju and Tapalang. (The author acknowledges valuable input from survey partner Thomas V. Laskowske to this report, especially in the determination of similarity sets and matrix relations. Maps la and 2a were drawn by Kare J. Str~mme.) The goal of the survey was to complete earlier UNHAS-SIL surveys in the area, i.e. the initial survey by Grimes and Grimes in January 1983, thus bringing the general survey of the whole area to completion. In other words our purpose was to investigate the linguistic situation in kabupaten Mamuju in order to establish language boundaries by comparing lexical similarity between wordlists taken in the area concerned. 1. R(MLIlIGOISTIC IlIPOlDfATIOR The area covered by our survey belongs to one administrative unit, kabupaten Mamuju. Under the Dutch rule and since independence until 1959 the area was a.subdivision of the Mandar administrative area. Kabupaten Mamuju is divided into six kecamatans: Tapa lang , Mamuju, Kalukku, Ka1umpang, Budong-Budong and Pasangkayu, counting from south to north. The whole area is still underdeveloped, the two northernmost kecamatans being the least developed. Large road building projects are under way. We were told that there is a plan to build a road along the coast from Mamuju to Bambamua in Central Sulawesi. At present there is a vehicular road from Mamuju to Benggaulu, but most of the bridges are still missing. We were told that 90 bridges are needed. There is also a road building project from Tasiu to Kalumpang. At present the road has reached Pabettengan, but again the bridges are missing. There are usually rafts to get motorcycles across the rivers and sometimes the rafts are big enough for a car. Until the roads and bridges are completed, most travel of any distance is by boat. There is regular traffic from Mamuju to Kalukku, Sampaga, Babana and Pasangkayu. One can also travel upstream on the big rivers: Karama, Budong-Budong and Lariang. If one can travel as an ordinary passenger, the fare is usually cheap, but he may need to wait for a boat that is leaving for his 100 MAMUJU destination. The other way is to charter a boat, but then one pays a lot more. Travelling by boat during the rainy season can be dangerous. Kabupaten Mamuju is basically mountainous, except for the coastal plain, which can be from five to 20 kilometers wide. Most of the people live on the coastal plain. The mountainous areas are, generally speaking, uninhabited, except for kecamatan Kalumpang. But even there most of the people live in the valleys. We were told that there are isolated groups living in the mountains in the Budong-Budong and Pasangkayu kecamatans, but it is difficult to verify whether that is true. The government has programmes to move those isolated groups down to more developed areas, where there are schools and clinics. There is a lot of migration going on in the area, especially in the sparsely populated Budong-Budong and Pasankayu kecamatans. There are both local and national transmigration projects that are being carried out and more are planned. This may be an important factor for the language, situation in the areas concerned. Subsistence farming is the livelihood of most people in the area, rice being the most ~portant crop. Both wet and dry field cultivation is used. Corn, cassava, soybean and sweet potato are cultivated to a lesser extent. Dried fish is part of the daily diet, chicken is eaten less frequently. Copra and coffee seem to be the most common cash crops. Cocoa and cloves are also cultivated and rattan is cut from the jungle. The majority of the villages have pr~ary schools (over 130 SDs in the whole kabupaten). There is a secondary school (SMP) in each kecamatan (two in Tapalang) and one high school (SMA) in Mamuju. There are over 100,000 inhabitants in kabupaten Mamuju. About 85% are Muslim, about 12% are Christian and about 1% are Hindu. 2. METHODOLOGY 2.A. General The goal of the survey was to determine tentative language boundaries by comparing lexical similarity between wordlists taken in the area concerned. Our definition of language ~plies mutual intelligibility. So, when we speak of two separate languages, we imply lack of intelligibility. Even though lexicostatistics is not a sufficient means of predicting intelligibility, we assume that the measure of lexical similarity roughly correlates with the degree of intelligibility. We feel that lexicostatistics can be used to indicate language relationships, but that any resultant classification must be considered tentative. (Smith 1982.) Dyen MAMUJU 101 notes: "The reason that 1exicostatistica1 classification is inconclusive would be the same that affects any classification; not all the knowable facts are known at the time of classification unless one chooses to wait hopelessly until all knowable facts are known. Based on fewer facts than that, a classification remains open to correction as additional facts become available." (Dyen 1966:35.) The additional facts that make the language classification final would be intelligibility testing between all those language communities which show lexical similarity between 60 to 95 percent. This is because, whereas lexical similarity is one component of linguistic similarity (together with phonological, grammatical and semantic factors), intelligibility scores represent linguistic similarity directly, being composite measures of the several factors which determine linguistic similarity. (Simons 1979:15,57f,67-87.) 2.B. Eliciting data During the survey trips 26 word1ists were elicited. All those word1ists are listed in Matrix 3 (see appendix). Lists elicited during the first trip are marked (L/V), those of the second trip are marked (B/V), and those of the third trip are marked (S/V). In the smaller South Sulawesi Stock matrix (Matrix 2) Pedasi and Tobada lists were not used because the former was an incomplete list (containing only 41 words) and the latter was elicited from an nonnative speaker. We also included eleven word1ists from Grimes and Grimes: Languages £f South Sulawesi (to appear), four word1ists from Str~mme and Valkama's survey of southwestern section of kabupaten Po1ewali Mamas a and four wordlists from Friberg and Kim's survey of the northern part of kecamatan Mambi in kabupaten Polewa1i Mamasa. These are marked (G), (S/V), and (F/K), respectively, in our matrix. We also included one word1ist from Michael Martens (M) and two word lists from Seko survey made by Laskowske and Laskowske (L/L). This makes the total number of word1ists 48. The wordlist we used was a 210-item list, which was quite similar to the Grimes and Grimes, Str~mme and Valkama, Friberg and Kim and Laskowske and Laskowske lists, and identical with the Martens list. Out of the 210 items, 200 were also on the Grimes and Grimes list, 190 were on the Str~mme and Va1kama list, 207 were on the Friberg and Kim list and 209 were on Laskowske and Laskowske list. All these lists are substantially similar to the Swadesh 200 list. Before eliciting a wordlist, the respondent was screened to ensure that he was a native speaker and the offspring of native speakers of the language/dialect in question. We permitted other local people to advise and correct the respondent, but we required 102 MAMUJU that the respondent hhnself pronounce the words, so that uniformity of pronounciation could be maintained for a given location. When Pedasi and Tobada wordlists are not counted, the following statistics can be given from the remaining 24 wordlists elicited during our Mamuju survey: Our respondents' ages ranged from 21 to 11: 2 were between 20 and 29, 8 were between 30 and 39, 4 were between 40 and 49, 7 were between 50 and 59, 1 was between 60 and 69, and 2 were between 70 and 79. 8 of the respondents were farmers, 1 was ketua adat, 1 was ketua kantor agama, 1 was karyawan CV, 1 was kepala RT, 1 was kepala RK, 1 was kepala lingkungan, 2 were kepala desas, 1 was kepala jaga, 1 was isteri RT, 1 was ketua BP3, 1 was sekretaris desa, 2 were imams, 2 were guru SDs. 21 of our wordlists were elicited in the respective speech community and 3 of them outside of the speech community. Only one of the respondents was female. The language of elicitation was Indonesian. 2.e. Lexical siailarity All 48 word lists were later transferred to a wordbook where each page contained one item in all its 48 varieties. Where a given wordlist elicited more than one response, each response was entered on the wordbook page. After that each word was assigned to its appropriate lexical similarity set according to the principles explained in Bugenhagen (1981:12-14). A lexical similarity set is a set of forms which are all lexically similar to one another (Sanders 1971:36). It is to be noted that here we use the terms lexically similar and lexical similarity set instead of the terms cognate and cognate set, because we want to make it clear that we are making synchronic comparisons rather than diachronic comparisons. Bugenhagen refers to McElhanon (1961:8), according to whom two items are lexically similar if there is correspondence between fifty percent or more of the phonemes, either as a regular correspondence (for the comparative method) or by phonetic similarity (for the inspection method). (For a discussion of comparative method and the inspection method, see Sanders 1971:33,34.) Bugenhagen combines the comparative method with the inspection method to the extent that regular correspondences are taken into consideration. However, he does include loanwords, which are excluded in the comparative method. So, basically we followed the inspection method, with the addition that regular correpondences were also taken into consideration. The decisions were made on a lexically similar/lexically different basis according to the 50% correspondence guideline. MAMUJU 103 2.D. Disqualifications Out of the 2I0-item wordlist, six items were disqualified because of ambiquity, lexical repetition, confusion, and lack of correspondence between the language of elicitation and the ll language elicited. The items were #39 "danau", #47 "hangat , 1105 "timur" ,11106 . "barat!", fJ:152 "nenek moyang", and fJ:207 "menjatuhkan". A few other items were disqualified from some wordlists, because we strongly suspect that the wrong item was elicited. So, after the disqualifications, the wordlist containes 206 items, and some individual wordlists even less. 2.E. Matrix A matrix containing the comparisons between the 48 word lists was then produced with the help of a computer. The procedure for "reducing" matrices as outlined by Gary Simons (1977:75-105) and as used in the UNHAS-SIL survey of southwestern Polewali-Mamasa and kabupaten Majene was not used here, for the following reason. Though we realize that two percentage figures, say 68% and 75% may not be significantly different: at the confidence level of .10, the representation of that in the matrices using the number of comparisons we used (1128 comparisons on 48 word1ists) becomes meaningless. The number of instance of percentages between 70 to 79 inclusive is represented below. percentage: 70 71 72 73 74 75 76 77 78 79 number of instances: 12 14 13 17 11 12 13 23 19 19 If we arbitrarily, for the sake of illustration, reduce 70%-74% to 72% and 75%-79% to 77% we are hard put to explain why a 69%-73% and a 74%-78% reduction to 71% and 76%, respectively, isn't equally valid. Furthermore, we hide the fact that two percentages, here 74% and 75%, are not significantly different while 70% and 79% are shown to be, though both sets are represented as the reduced pair, 72% and 77%. It seems best to abandon a matrix representation of reduction for as many comparisons as we have been using. This abandonment notwithstanding, we still showed sensitivity to significant differences in our assignment of the accompanying linguistic taxonomy. 104 MAMUJU 3. HEIUL'lS 3.A. Gelleral Percentages above 80 are considered to indicate one language. Percentages between 75 and 80 are considered to indicate the same subfamily of languages; between 60 and 75, the same family of languages; between 45 4nd 60, the same stock of families; between 25 and 45, the same superstockof stocks. (Gr~es and Grimes, Languages of South Sulawesi (to appear).) All the wordlists of this survey belong to the same West Austronesian Superstock. (See Matrix 3.) Lexical similarity percentages are generally not used to differentiate between dialects, 80 we did not break languages down into dialects based on that. 3.B. Central Sulawesi Stock As the Central Sulawesi Stock matrix shows (Matrix 1.), this stock i, divided into two families: the Kaili-Pamona Family, which contains the Uma, Sarudu, Kaili, and Topoiyo languages, and the Bada Family, of which Ako· is a member. Matrix 1. Central Sulawesi Stock .Bada famill- -Kaili-Pamona Uma Family Sarudu Kaili 3.B.1 The BacIal'alBily The inhabitants of the Ako' village cla~ to speak the Bada language of Central Sulawesi, and our Aka' wordlist when compared with the laO-word Bada list found in Barr, Barr and Salombe, shows 90% lexical similarity with it. According to them, they moved from Bada between 100 and 150 years ago. The high lexical similarity with the Bada list poses a problem, however. According to Barr, Barr and Salombe, the Bada language is a member of the Kaili-Pamona Family, but according to our matrix, the language MAMUJU 105 spoken in the village of Ako' is clearly a member of a different family, which we call the Bada Family, which according to our calculations has a status independent of the putative Kaili-Pamona family of Barr, Barr and Salombe (aBBS). A look at the comparisons may help clarify what we are saying. BBS matrix BBS word lists This survey's counted by Valkama .wordlist Bada Bada Ako' -~ ~ IUma Uma Uma Da'a 65 Da'a Da'a ~ [ill 1 ~ Clearly there is a discrepancy in calculation procedures. Though this is not the place to resolve this difference, we may note that figures listed in Grimes and Grimes for Topoiyo, Sarudu, Pamona and Rampi lend support to our separation of Bada, showing that several of the putative members of the Kaili-Pamona Family relate to each other, not as members of the same family, but only as members of the same Central Sulawesi Stock. (e.g., Topoiyo and Rampi relate to each othep at only 34%.) More research is assuredly called for. 3.B.2 The Kaili Pamona Fa.ily As noted immediately preceding, the family status of Kaili-Pamona is in doubt. Furthermore, since Sarudu relates both to Uma and Kaili at an average of over 75% lexical similarity (75.3% and 75.6% respectively), we have not divided the family into subfamilies. The inhabitants of Benggaulu claim to have come from Kantewu 1n Central Sulawesi, therefore it is natural that Benggaulu shows 90% lexical similarity with Kantewu. The Sarudu language is closely related to both Uma and Kaili. The Grimeses' wordlist is closer to Kaili and our wordlist is closer to Uma. The Grimeses' wordlist shows only 86% lexical similarity with our Nunu list. We were unable to trace where the Grimeses' respondent came from, most probably from a village geographically close to the Kaili Family. All the villages (about 11) of desa Sarudu speak Sarudu, so the language seems quite stable. The Kaili language is represented by four dialects 1n our matrix: 1. Da'a/lnde, known also as Bunggu/Binggi in kecamatan Pasangkayu. This dialect is represented by the Dombu and Tosande I word lists in our matrix. The main area of this dialect is in Central Sulawesi. 2. Tado, represented by the Kabuyu wordlist. According to our information, this dialect is spoken only in desa Pasangkayu. 3. Baras, represented by the Salubiro and Bambaloka 106 MAMUJU word1ists. This dialect is spoken only in a few of the villages of desa Baras, by about 50 households. The elders fear that their language is dying out, which remains to be seen. 4. Doi, which is represented by the Tampaure wordlist. Three villages in desa Bambaira spoke this dialect. The main area of this dialect is in Central Sulawesi. Also the Ledo, Unde and Rai dialects are spoken in kecamatan Pasangkayu. The main area of those dialects is in Central Sua1wesi, so we did not take wordlists for them. The Topoiyo language 'was originally spoken by only one village on the Budong-Budong river in kecamatan Budong-Budong. At present it has spread to the neighbouring Tobada village, where it was claimed that only one speaker still spoke the Tobada language. It is also spoken in the Tangkou village. So, even though a small language, it seems to be vigorous and not on the edge of extinction. 3.C. South Sulawesi Stock The South Sulawesi Stock is represented by the Seko Family and the Northern South Sulawesi Family. (See Matrix 2.) 3.C.l. The Seko Family Panasuan is spoken 1n two villages in kecamatan Kalumpang and Tangkou is spoken in the Tangkou village on the Budong-Budong river in kecamatan Budong-Budong. It is also called the Budong-Budong language, which was included as a dialect of Mamuju in Grimes and Grimes' Languages of South Sulawesi (to appear). Tangkou is spoken only by about 11 households. Intermarriage and proxi~ity with Topoiyo seems to be taking over. To more firmly establish Tangkou as a member of the Seko Family, it has to be compared with Seko Padang and Seko Tengah. 3.C.2. The Borthern South Sulawesi Pamily The Northern South Sulawesi Family is represented in our matrix by the Toraja and the pitu Ulunna Sa1u (PUS) subfamilies and the Mamuju and Mandar languages, which do not belong to any subfamily. The Toraja subfamily is represented by five languages in our matrix, of which Talondo' and Kalumpang are spoken in kabupaten Mamuju. 3.C.2.l. The Toraja Subfamily The Talondo' language is spoken in one village, Ta10ndo', in desa Bonehau in kecamatan Ka1umpamg. It shows at an average 80% lexical similarity with the Kalumpang language, and was separated from that language because all Kalumpang speakers with whom we discussed the Ta1ondo' language, said that they cannot understand it. The Ta1ondo' list's high lexical similarity with Tarata, Pabettengan and Ka1umpang is explained by proximity and by the MAMUJU 107 Matrix 2. South Sulawesi Stock 108 fact that lexical similarity gives only an initial picture of the language situation. Intelligibility testing is needed to give the final word. The Kalumpang language is divided into eight dialects according to Mr. Darius Timbonga, who himself speaks Panasuan as his mother tongue, but he lives at present in kota Kalumpang. He seemed to be very informed about the dialect variations in kecamatan Kalumpang. It is to be noted, however, that, according to him, desa Karama, which is his home desa, is divided into many small dialects, while dialects in other desas tend to be larger. This shows that he has best information from his own desa, and tends to divide even slightly different speech forms into separate dialects. It remains to be seen whether there are more dialects in other desas, too, or whether some of his dialects in desa Karama should be grouped together. According to him the dialect situation is as follows (underlining indicates a village from where a word1ist has been included in out matrix): 1. Mariri and Limbong villages. 2. Petangunan and Buttu villages. 3. Tambing-Tambing and Kondoan villages. 4. The Te'da dialect, spoken in the Sabamba, Tararang, Sa1ukayu, Rantepata, Sa1upolin and Bau villages. 5. Bu~l10, Pekkaro and Salu villages. 6. The E'da dialect, spoken in the Lebani, Bulo, Sa1usokang, Malo10, Tala, Lasa, Batuisi, Henua, Salu Banga, Salu Eno, Pelosian, Paradang, Sa1utake, Ka11an Baru and Salu Batu villages. (Lasa is mixed with the Toraja language.) 7. The Ta'da dialect, spoken in Tamalea, Pabettengan, Lossa, Tama1eatua, Sumua, Kalumpanf' Tamessassang, Tarata, Tamanggeso, and Tambuku. Some people claun that Ta'da should be divided into three dialects: Sandana (Tarata and Tamessassang), Leling (Tambuku and Tamangeso) and the rest. We leave this question open. It is to be noted, however, that this is the reason that the Tarata list has Sandana/Ta'da tag in the matrix. 8. Dengen, Mawunggin and Tadossan villages. The Ta'da dialect is also called the Bone Hau dialect and the E'da dialect is also called the Karataun or Makki/Mangki dialect. These two dialects are the biggest and most important Ka1umpang dialects. Grimeses' Ka1umpang respondents were from Bu10 in desa Karataun and Rantepata in desa Karama. The man from Bu10 told us that he had mixed several dialects when he gave the word1ist. It is therefore certain that it does not represent the speech form used in kota Ka1umpang. This also explains why our Kalumpang list shows only 85% lexical similarity with the Grimeses' Kalumpang list. 3.C~2.2. The Pitu Dianna Salu Subfamily Out of the pitu U1unna Sa1u subfamily, only the Ulumandak language is spoken in kabupaten Mamuju. It is also spoken 1n kabupatens Majene and Polewa1i Mamasa. Later, one matrix covering the whole area is needed. According to our findings, Botteng, MAMUJU 109 Tasiu (Sondoang), and Orobatu (Tappalang) belong to the Ulumandak language as opposed to the Grimeses' grouping of them as dialects of Mamuju. Van der Veen (1929:81-86) includes the area covered by our Ulumandak language as dialects of Pitu Ulunna Salu. According to him the speech form, which we call Ulumandak, is spoken in Kaluku, Rante Dango, Padang Baka, Padang Panga, Taludu and most of the "landschap Talapang" (Talapang area), for example in Lebani, and Lombang-Lombang. Landschap Talapang seems to correspond with kecamatan Tapalang. Generally speaking van der Veen's article is accurate and conforms to a large extent to the findings of our survey. Ortobatu shows an average of 82.6% lexical similarity with the rest of Ulumandak and an average of 80.3% lexical similarity with Mamuju. With Ulumandak its lexical similarity percentages are consistently above 80% only with Padang, Sumare and Rangas. Its high lexical similarity with Padang, Sumare and Rangas can be explained by proximity and contact more easily than the consistently high percentages with Ulumandak. 3. C. 2.3. 'ftle Mamuju language The number of dialects of Mamuju stated in Grimes and Grimes' Languages of South Sulawesi has been reduced by five: Botteng, Sondoang CTasiu), Tappalang (Orobatu) and Budong-Budong have been assigned to other languages while Sumau and Rangas have been shown to be the same dialect. What remains then are the dialects of Mamuju (Mamuju, Babana and Sampaga), Sumare/Rangas, Padang (Anjoro Pitu) and Sinyonyoi (Tampapadang). We have nothing to say against that dialect grouping, although Grimes and Grimes' location of these dialects on their map is inaccurate. 4. RESIDUE There are still several questions that need more study: 1. The relationship between the Bada Family (Ako' language) and the putative Kaili-Pamona Family. 2. The relationship of Talondo' to Kalumpang. Is it separate language or not? 3. The border between the Ulumandak and Mamuju languages. Both Orobatu and Anjoro pitu show lexical similarity at an average above 80% with both languages. 110 MAMUJU BDLIOGUPBY Bugenhagen, Robert. 1981. A Guide for Conducting sociolinguistic Surveys in Papua New Guinea. Summer Institute of Linguistics. Dyen, Isodore. 1966. Comment. Oceanic Linguistics 5:1.32-49. Friberg, Timothy. UNHAS-SIL sociolinguistic survey: Kabupaten Polewali Mamasa, Northern Section. In this volume. Grimes, Charles E. and Barbara D. Grimes. 1984. Languages of South Sulawesi. To appear. Laskowske, Thomas V. UNHAS-SIL sociolinguistic Survey: Seko Area. In this volume. Loving, Richard. 1977 (ed.). Language Variation and Survey Techniques. Workpapers in Papua New Guinea Languages, Vol. 21. ----. Mamuju Dalam Angka 1981/82. Kantor Statistik Kabupaten Mamuju. McElhanon, K. 1967. Preliminary Observations on Huon Peninsula Languages. Oceanic Linguistics 6:1-45. Pelenkahu, R.A., et ale 1974 (eds.). Peta Bahasa Sulawesi Selatan. Ujung Pandang: Lambaga Bahasa Nasional Cabang III. Salombe, C., Don Barr and Sharon Barr. 1979. Languages of Central Sulawesi. Hasanuddin University, Ujung Pandang. Sanders, Arden G. 1977. Guidelines for Conducting a Lexicostatistic Survey in Papua New Guinea. In Loving ed., 1977. Simons, Gary F. 1977. Tables of significance for lexicostatistics. In Loving. Simons, Gary F. 1979. Language Variation and Limits to Communication. Technical Report No.3. Ithaca: Cornell University. Smith, Kenneth D. 1982 (ed.). A Compendium of Articles Relating to the Survey of Sabah Languages. To appear. MAMUJU 111 Veen, H. van der. 1929. Nota Betreffende de grenzen van de Sa'dansche taa1groep en het aanverwante taalgebied. Tijdschrift voor Indische Taa1-, Land- en Volkenkunde 69:58-97. 112 MAMUJU llatru 3 - West Anatronesian Superstock .- MAMUJU 113 Map 1a - Eabupaten ~ju, Borthern Section . l~a.tA L.~e"J (S~ •.I. I: S"oo 1100/: ~ Proyine-ferd bo,.der 114 MAMUJU Hap Ib - LanKUAcea of Eabapaten Ha.aju, Borthern Section MAMUJU 115 Map 2a Kabupaten Ma.uju, Southern Section 116 MAMUJU Hap 2b - L=&n&cea of Ealnapaten Kamju, Southern Section MAMUJU 117 page 118 [blank] URBAS-SIL SOCIOLIHGUISTIC SURVEY: XABUPATERS PDlRARG, ENREKABG, TARA roRAJA, LUWU AND EASTERN PAKT OF POLEWALI MAHASA Kari Valkama TABLE OF COHTERTS Page INTRODUCTION 120 1. NONLINGUISTIC INFORMATION 120 2. METHODOLOGY 121 3. RESULTS 123 4. RESIDUE 131 S. BIBLIOGRAPHY 131 MATRIXES: Matrix 1. Mamasa and Toraja 126 Matrix 2. Luwu 127 Matrix 3. Masenrempulu 129 Matrix 4. Central Sulawesi Stock 130 APPENDICES: A. Matrix S. South Sulawesi Stock 133 B. Matrix 6. Central Sulawesi Stock 134 C. MAPS: South Sulawesi Map 134 Map 1. Northcentral South Sulawesi Map 135 Map 2. Northeastern South Sulawesi Map 136 TORAJA 119 IlITRODUCTlOR This is a report of the Masenrempulu, Toraja and Luwu survey trips. The first trip, made by Juha Christensen and Kari Valkama April 15 - 24, 1986, covered the northern part of kabupaten Pinrang, kabupaten Enrekang and the eastern part of kabupaten Tana Toraja. The second trip, made by Juha Christensen and Kari Valkama May 25 - June 7, 1986, covered the northern and eastern part of kabupaten Luwu. The third trip by Philip Campbell and Kari Valkama August 20 - September 2, 1986, covered the eastern part of kabupaten Polewali Mamas a and the western part of kabupaten Tana Toraja. The fourth trip by Juha Christensen and Kari Valkama September 21 - 27, 1986, covered the southern part of kabupaten Luwu. The goal of the survey was to investigate the linguistic situation in kabupatens Pinrang, Enrekang, Tana Toraja, Polewali Mamas a and Luwu in order to establish language boundaries by comparing lexical similarity between wordlists taken in the areas concerned. This was to complete the earlier UNHAS-SIL survey (i.e. the initial survey by Grimes and Grimes in 1982/1983) in the areas mentioned above, thus bringing the general survey of the area to completion. 1. ROlILIBGIJISTlC IRFOIDIATlOR The area covered by our survey belongs to five administrative units, kabupatens Pinrang, Enrekang, Tana Toraja, Polewali Mamasa and Luwu. In Pinrang we visited kecamatans Patampanua, Duampanua and Suppirang. In Enrekang we visited all five kecamatans: Maiwa, Enrekang, Baraka, Anggeraja and AlIa. In Tana Toraja we visited kecamatans Mengkendek, Makale, Sanggalangi, Sesean, Rantepao, Rindingallo, Saluputti and Bonggakaradeng. tn Polewali Mamasa we visited kecamatans Mamas a, Sumarorong and Polewali. tn Luwu we visited all the kecamatans except kecamatan Lifubong. See Appendix C for map of area surveyed. The condition of the roads is basically good. Most of the roads were asphalt. In Pinrang we had some difficulties when going to desa Kassa and towards desa Suppirang. Those roads were passable by motorbike only during the dry season. Kecamatan Pana in kabupaten Polewali Mamas a , kecamatan Bonggakaradeng in kabupaten Tana Toraja and kecamatan Bastem in kabupaten Luwu were not accessible by motorbike. The area concerned is basically mountainous, except for the coastal plain of Pinrang and the southern part of kecamatan Maiwa and the coastal areas of kecamatans Polewali, Mamasa and Luwu. Farming is the livelihood of most people in the area, rice being the most important crop. Wet field cultivation is most common. 120 TORAJA Corn and cassava (ubi kayu) are cultivated to a lesser extent. Animal husbandry is also important in many areas. There are over 100,000 inhabitants in the three above-mentioned kecamatans of Pinrang. About 98% are Muslim and about 2% are Christian. In kabupaten Enrekang there are over 130,000 inhabitants out of which about 1% are Christians and the rest are Muslim. In kabupaten Tana Toraja there are about 340,000 inhabitants. About 80% are Christians, 7% Muslim and 13% an~ists. In kabupaten Polewali Mamas a in kecamatans Mamasa, Pana, Sumarorong and Polewa1i there are about 130,000 inhabitants. About 56% are Muslim, the majority of whom live in kecamatan Polewali, 40% are Christians, most of whom live in kecamatan Mamasa and Pana, and about 4% are an~ists. In kabupaten Luwu there are 560,000 inhabitants, the majority of which are Muslim. 2. HBTIIODOLOGY 2.A. General The goal of the survey was to determine tentative language boundaries by comparing ,lexical similarity between wordlists taken in the area concerned. Our definition of language ~plies mutual intelligibility, so, when we speak of two separate languages, we imply lack of intelligibility. Even though lexicostatistics is not a sufficient means of predicting intelligibility, we assume that the measure of lexical similarity roughly correlates with the degree of intelligibility. We feel that lexicostatistics' can be used to indicate language relationships, but that any resultant classification must be considered tentative (Smith 1982). Oyen notes: "The reason that lexicostatistical classification is inconclusive would be the same ,that affects any classification; not all the knowable facts are known at the time of classification unless one chooses to wait hopelessly until all knowable facts are known. Based on fewer facts than that, a classification remains open to correction as additional facts become available." (Oyen 1966:35). An additional factor that would make the language classification final would be intelligibility testing between all those language communities which show lexical similarity between 60 and 95 percent. This is because whereas lexical similarity is one component of linguistic similarity (together with the phonological, grammatical and semantic components) intelligibility scores represent linguistic similarity directly, being composite measures of those components which determine linguistic similarity (S~ons 1979:15,57f, 67-87). TORAJA 121 2.D. Eliciting data During the four survey trips, 65 wordlists were elicited. 60 of those wordlists and 19 wordlists from other surveys were compared. They are listed in two matrixes: Matrix 5, Appendix A (63 wordlists) and Matrix 6, Appendix B (16 wordlists). Lists elicited by the Grimeses are marked (G), lists elicited by the Fribergs are marked (F), lists elicited by the Laskowskes are marked (L), lists taken from Barr & Barr are marked (B) and a list taken from Balai Penelitian Bahasa is marked (BPB). This makes the total number of wordlists 79. The wordlist we used was a 2l0-item list, which was quite similar to the Grimeses, Friberg and the Laskowske lists. Barr and 'Barr used a modified Swadesh 100 list and Balai Bahasa used the Swadesh 200 list. Out of the 210 items, 200 were also on the Grimes list, 210 were on the Friberg list and 209 were on the Laskowske list. All these lists are substantially similar to the Swadesh 200 list. Before eliciting a wordlist, the respondent was screened to ensure that he was a native speaker of the language/dialect in question. We permitted other local people to advise and correct the respondent, but we required the respondent himself to pronounce the words, so that uniformity of pronounciation could be maintained for a given location. The language of elicitation was Indonesian. 2.e. Lexical si.ilarity All 79 word lists were later transferred to a wordbook where each page contained one item in all its 79 varieties. Where a given wordlist elicited more than one response, each response was entered on the wordbook page. After that each word was assigned to its appropriate lexical similarity set according to the principles explained in Bugenhagen (1981:12-14). A lexical similarity set is a set of forms which are all lexically similar to one another (Sanders 1977:36). It is to be noted that here we use the terms lexically similar and lexical similarity set instead of the terms cognate and cognate set, because we want to make it clear that we are making synchronic comparisons instead of diachronic comparisons. Bugenhagen refers to McElhanon (1967:8), according to whom two items are lexically similar if there is correspondence between fifty percent or more of the phonemes, either as a regular correspondence (for the comparative method) or by phonetic similarity (for the inspection method). (For the discussion of comparative method and inspection method, see Sanders 1977:33,34.) Bugenhagen combines the comparative method with the inspection method to the extent that regular correspondences are taken into 122 TORAJA consideration. Loanwords, however, he leaves in, while in the comparative method they are left out. So, basically we followed the inspection method, with the addition that regular correspondences were also taken into consideration. The decisions were made on a lexically similar/lexically different basis according to the 50% correspondence guideline. 2.D. Disqualifications Out of the 210-item word1ist, 13 items were disqualified because of ambiguity, confusion, lexical repetition, lack of correspondence between the language of elicitation and the language elicited and difficulties encountered in assigning words into their lexical similarity sets. The items were 1126 "mer eka'", i139"danau", 4147"hangat '", 4194 11 ini", #95 "itu", 197 "di situ", 11106 "barat". " 1F149 "paman" 1152 "nenek moyang" , #156 "emas kawin", i1157 "j eluang", iH70 "di mana", and 11205 "mandikan". 2.E. Matrix Two matrixes containing the comparisons between the 79 wordlists were then produced with the help of a computer. The larger one (Matrix 5, Appendix A) contains South Sulawesi Stock wordlists. The smaller one (Matrix 6, Appendix B) contains Central Sulawesi Stock word1ists. The larger matrix was then split into three smaller matrixes: Toraja, Luwu and Masenrempu1u matrixes (matrices 1, 2 and 3, respectively). In the smaller matrixes some of the word lists were disqualified because we suspect that they do not represent the respective speech form. Therefore the Pattae' wordlist was disqualified because it has a Bugis bias. The Ba'atan wordlist was taken from a person who was born in the Tallulembangna dialect area but later moved to the Kesu' dialect area. Therefore it is a mixed wordlist. The Kariango wordlist was disqualified because the quality of the elicitation was thought to be poor. The Bungi and Maroangin wordlists also showed a Bugis bias. 3. RESULTS 3.A. General Percentages above 80 are considered to indicate one language. Percentages above 75 are considered to indicate the same subfamily of languages. Percentages between 60 and 75 are considered to indicate the same family of languages. Percentages between 45 and 60 are considered to indicate the same stock of families. Percentages between 25 and 45 are considered to indicate the same TORAJA 123 superstock of stocks. All the wordlists of this survey belong to the South Sulawesi Stock or the Central Sulawesi Stock. Lexical similarity percentages are generally not used to differentiate between dialects, so we relied more on sociolinguistic factors in breaking languages into dialects. It is to be noted that we did not follow the percentages blindly when separating languages from each other. We tried to take sociolinguistic factors into consideration also. It is also to be remembered that lexicostatistics gives only an approximation of the real language situation. To finalize the picture, thorough fieldwork and intelligibility testing is needed. We are in no way trying to imply that this survey report presents the final and true relationship between the languages concerned. Two other languages treated in Grimes and Grimes, Rampi and Lemolang, are found in the area of our survey, namely in kabupaten Luwu. We have not included these because they are so unrelated to everything else. Lemo1ang stands as a language isolate within the South Sulawesi Stock. Rampi is assigned the Central Sulawesi Stock. We have nothing to add beyond Grimes and Grimes and the Rampi survey (in this volume). 3.B. South Sulawesi Stock 3.B.1. Bugi. Fa.ily As Matrix 3 shows, this family is represented by Bugis and Malimpung wordlists. The Bugis list represents the Sidrap dialect and therefore (because of close location) some of the other wordlists show over 60% lexical similarity with it. (See Maiwa, Pattae' and Pattinjo-Enrekang wordlists.) 3.B.2. Rorthern South Sulawesi Pa.ily 3.B.2.1. Toraja Subfa.ily This subfamily is represented by the Toraja, Mamasa, (Matrix 1) and Luwu (Matrix 2) languages in our matrix. The Toraja language is represented by the Sa'dan, Kesu', Tallulembangna and Saluputti-Rindingallo dialects. Duri relates to Toraja dialects with an average of 77.8% lexical similarity and to Enrekang-Pattinjo with an average of 75.7%, so it could have been included in the Toraja subfamily. But because of sociolinguistic factors, we included it in the Masenrempulu subfamily. This is in accordance with van der Veen, Peta Bahasa and Grimes and Grimes. The Mamas a language is represented by the northern dialect (kecamatan Mamas a and Pana), the middle dialect (kecamatan Sumarorong, desas Mappak and Simbuang in kecamatan Bonggakaradeng and desa Suppirang in kecamatan Lembang) and the southern or Pattae' dialect (kecamatan Polewali). According to the Grimeses, Pattae' is a separate language from Mamasa. Their only wordlist of Pattae' is from Binuang, 124 TORAJA which is heavily influenced by Bugis. In our matrix it is 11-12% closer to Bugis than our other Pattae' wordlists. Therefore we have reason to believe that that word list does not represent a pure Pattae' dialect. Therefore we excluded that list from Matrix 1. Since Pattae' relates to other Mamasa dialects with an average of 86.7% lexical similarity, it is clearly part of the Mamasa language even though there are some sociolinguistic factors that may cause Pattae' speakers to reject materials written in the northern dialect. This decision is in accordance with van der Veen and Peta Bahasa. The northern and middle dialects of Mamasa relate to Toraja with an average of 84.9% lexical similarity. It is almost as much as the percentage between Pattae' and the two other Mamasa dialects. We decided not to include Mamasa in the Toraja language because of the following factors. The deviation of the percentage figures between Toraja and Mamasa (excluding the Pattae' dialect) is higher than the deviation of percentage figures between Pattae' and Mamasa (the Mamasa average excludes the Pattae' dialect). We believe that this fact is the result of contact between Mamaaa and Toraja which has resulted in bilingualism for those language communities that are geographically close to Mamasa. For example, our Balla list (which is geographically close to Mamasa) shows an average of 90% lexical similarity with Mamasa (excluding the Pattae' dialect), but our Bonoran list, which represents the prestigeous Kesul dialect (and which is geographically distant from Mamasa) shows an average of only 80,6% lexical similarity with Mamasa (excluding the Pattae' dialect). All the Pattae' wordlists, on the contrary, show an average of 85% or more lexical similarity with the Mamas a (the Mamas a average excludes the Pattae' dialect). The Luwu language is a hard one to define (see Matrix 2). It is represented by Rongkong, Luwu Utara and Luwu Selatan dialects. (These can be divided into smaller units, like Rongkong Bawah, Rongkong Atas, Seko Lemo, northern Luwu Selatan, southern Luwu Selatan, Sabbang and Bone Bone dialects.) In the people's minds Rongkong is separate from the Luwu language. But according to our lexcial similarity figures, Rongkong is, on the average, 88,9% lexically similar with Luwu Utara and on the average, only 82,4% lexically similar with Luwu Selatan. Luwu Utara and Luwu Selatan are on the average 85% lexically similar. So, according to the matrix, Rongkong is closer to Luwu Utara thanLuwu Utara is with Luwu Selatan. Therefore, if a split should be made, Luwu Utara should be grouped together with Rongkong and Luwu Selatan should be left alone. This split would have been against the intuition of the population, so we did not do it. TORAJA 125 126 TORAJA TORAJA 127 What is clear, however, is that Luwu Utara is the central dialect and that it's Sabbang dialect (represented by our Lena wordlist) is the linguistic center for the area. It remains to be seen if speakers of the other dialects can understand and are willing to acquiesce to the centrality of this dialect. Grimes and Grimes' Toala' is the same as our Luwu. Whereas they divide the language into two dialects, Toala' and Palili', on a north-south axis (which incidentally continues north and demarcates the Atas and Bawah dialects of the Rongkong language), we have spoken of northern and southern dialects based on a east-west axis. Both are correct divisions in th~ south. (In the north this division is corre~t only if Rongkong is considered a part of Luwu, a position that Grimes and Grimes do not take.) Thus Ranteballa, Paragosi, Bonelemo and Kaili (southern and western (mountain) villages) form a group which is over 90% lexically similar. Gr~es and Grimes Palili' wordlist taken in Palopo. is about 96% lexically similar with our Kandoa wordlist, making it part of the Luwu Selatan dialect. 3.B.2.2. Massenrempulu Subfamily The Masenrempulu subfamily is represented by the Enrekang-Pattinjo, Duri and Maiwa languages (see Matrix 3). The Enrekang-Pattinjo language is represented by the Pattinjo and Enrekang dialects. The Guzi, Kassa, Bungi, Paku and Rampusa wordlists are usually referred to by the cover term Pattinjo. They are spoken in kabupatens Pinrang and Polewali Mamasa. Enrekang-Pattinjo reportedly also has a Ranga dialect in desa Ranga kecamatan Enrekang. We do not have a word list from that dialect. Grimeses grouped Duri, Enrekang and Maiwa into one language and Pattinjo as a separate language. Our wordlists, however; show that Duri, Enrekang and Maiwa are separate languages. And since Pattinjo relates at an average 84.1% lexical similarity with Enrekang, it was included in the Enrekang-Pattinjo language. This is in accordance with the Peta Bahasa. The Maiwa language is represented by the Maroangin, Mataka1i and Bungin dialects. Matakali is a transitional dialect between Enrekang-Pattinjo and Maiwa. Maiwa is also reportedly spoken in kabupaten Sidrap in kecamatan Duapitue, desas Bulucenrana, Batu, Compong, Betao and Balawae, and in kecamatan Pancarijang in desa Kulo. We do not have wordlists from those speech communities. The Duri language seems to be uniform, without great dialect differences. It is close to Toraja showing an average 77.8% lexical similarity with it, compared to 75.7% lexical similarity with Enrekang-Pattinjo. But because of sociolinguistic reasons, we have included it in the Masenrempulu subfamily. This is in accordance with Van der Veen and Peta Bahasa. 128 TORAJA Matrix 3 - Hasaenrempulu - 14 wordliat:•• BAROKO 94 BlLAJIN DURI LANGUAGE 95 96 CAKKE 94 94 97 BARAKA 74 75 75 75 RAMPUSA 72 72 74 73 86 PAKU 75 75 77 76 86 90 KASSA ENREKANG-PATTINJO LANGUAGE 79 77 80 80 90 87 89 GUZI 75 75 78 76 80 80 87 87 ENREKANG 74 74 78 76 81 83 87 88 97 ENREKANG (G) 74 74 75 74 79 82 82 84 83 86 MATAKALI MAIWA LANGUAGE 77 77 77 77 76 80 79 79 76 78 86 BUNGIN 63 64 67 65 69 74 77 71 68 70 71 69 MALIMPUNG (F) MALIMPUNG LANG. 58 58 59 58 64 71 69 66 60 63 65 65 81 BUGIS (F) BUGrS LANGUAGE 3.C. Central Sulawesi Stock 3.C.l. Mori Family 3.C.l.l. Bungku subfamily The Bungku subfamily is represented by two languges, Bungku and Menui. Both of those wordlists are from Barr and Barr. In their matrix they relate to each other at 85%. In our matrix they relate to each other at 77%. When we compared the Barr and Barr word1ists, which are in our matrix, with each other, our results averaged 8.8% lower than Barr and Barr's results. It seems apparent that they used different criteria when making their lexical similarity decisions. 3.C.l.1. Mori subfamily The Mori subfamily is represented by three languages, Mori Bawah (also called East Mori) , Mori Atas (also called West Mori) and Padoe (also called South Mori). Mori has five dialects (according to Ibu Pendeta Ayub Lakaoni): Molongkuni (our Babopada), Toroda (our Beteleme), Topimp0, Tobatu and Ngusungbatu. We had no way of verifying that information. The dialect spoken in Soroako and Nuha is also called the Soroako dialect. Babopada is included in the Mori Bawah language although it is less than 80% lexically similar with the other dialects. This is because the Babopada word1ist was taken outside of the main area of the dialect and we have only one wordlist, which may contain errors. We think it is safer to give it a dialect status rather than a language status in this initial stage. The Kawatak wordlist is a Karonsie dialect list taken in the Kawatak village. It is transitional between Mori Bawah and Padoe. It was included in the Padoe language. TORAJA 129 The Koropansu list represents the Tambe'e dialect of the Mori Atas language. J.C.2. Pamona Family The Pamona Family is represented only by the Pamona Tomoni dialect. One of the wordlists was taken by the Grimeses and the other one was taken by us. They should represent different dialects, but they are quite close. During the survey we were told that the Pamona Tomoni dialect is very homogenous. Pamona Tomoni is only 45.5% lexically similar with the Central Sulawesi Stock lists of our matrix, so it may be that it should be regarded as a separate Stock. J.C.J. Tolaki Family The Tolaki Family is represented by only the Tolaki language. That list is taken from a Balai Penelitian Bahasa book. They did not say from which village their lists were taken. The only reason we put it into our matrix was because we wanted to know how close it is to the neighbouring Bungku and Mori subfamilies. It is 53.3% lexically similar with the South Sulawesi Stock. Since we do not know how close the other Southeast Sulawesi languages are with the Central Sulawesi Stock, we do not know if it really belongs to Central Sulawesi Stock or not. Matrix 4 - Central Sulawesi Stock - 16 wordlists IMENUI (B) MENUI LANGUAGE 77rBUNGKU (B) BUNGKU LANGUAGE 55 58 BABOPADA 59 64 80 BETELEME 62 68 79 92 MORIBAWAH (B) MORI BAWAH LANGUAGE 64 68 76 90 93 NURA 63 68 76 91 92 99 SOROAKO 59 62 77 82 80 86 87 KAWATAK 55 60 75 78 78 79 80 85 WAWANDOLA (G) PADOE LANGUAGE 57 60 76 77 76 80 80 84 94 PABETA 59 61 76 77 76 80 80 84 94 96 KAWATA· 53 56 79 79 75 75 76 80 79 78 80 KOROPANSU MORI ATAS LANGUAGE 56 60 79 75 79 76 75 77 80 80 80 82 MORIATAS (B) 55 59 53 48 54 50 51 55 57 58 56 57 6llTOLAKI (BPB) TOLAKI LANG. 45 42 43 45 48 44 44 44 45 44 44 45 51 43 MANGKUTANA PAMONA 44 41 42 44 47 44 45 45 44 44 43 46 48 42 94 MALEKU (G) LANG. 130 TORATA 4. llESIDUE There are still several questions that need more study: 1. Why the percentages are so different between the Barr and Barr matrix and our matrix? 2. Does Kawatak belong to Mori Bawah or Padoe? 3. Is Babopada a dialect of Mori Bawah or its own language? 4. Does Pamona belong to the Central Sulawesi Stock or to another stock? 5. Does Tolaki belong to the Central Sulawesi Stock or to a Southeast Sulawesi Stock. _S. BIBLIOGIlAPIlY Bugenhagen, Robert, 1981. A Guide for Conducting sociolinguistic Surveys in Papua New Guinea. Summer Institute of Linguistics. Dyen, Isidore, 1966. Comment. Oceanic Linguistics 5:1.32-49. Grimes, Charles E. and Barbara D. Grimes to appear. Languages of South Sulawesi. Kantor Statistik Ujung Pandang. 1984. Enrekang Dalam Angka ---- 1984. Luwu Dalam Angka. ---- 1984. Pinrang Dalam Angka. ---- 1984. Polewali Mamasa Dalam Angka. ---- 1984. Tana Toraja Dalam Angka. Laskowske, Thomas V. and Kathryn B. Laskowske, in this volume. UNHAS-SIL. Sociolinguistic Survey. Seko Area. Loving, Richard, ed. 1977. Language Variation and Survey Techniques. Workpapers in Papua New Guinea Languages, Vol. 21. McElhanon, K. 1967. Preliminary Observations on Huon Peninsula Languages. In Oceanic Linguistics 6:1-45. Pelenkahu, R. A. eds. 1974. Peta Bahasa Sulawesi Selatan. Ujung Pandang: Lembaga Bahasa National Cabang Ill. Salombe, C., Don Barr and Sharon Barr. 1979. Languages of Central Sulawesi. Hasanuddin University. Ujung Pandang. TORAJA 131 Sanders, Arden G. 1977. ~uidelines for Conducting a Lexicostatistic Survey in papua New Guinea. In Loving, ed. 1977. Simons, Gary F. 1977. Tables of significance for lexicostatistics. In Loving, ed. 1977. Simons, Gary F. 1979. Language Variation and Limits to Communication. Technical Report No.3. Ithaca: Cornell University. Smith, Kenneth D., ed. 1982. A Compendium of Articles Relating to the Survey of Sabah Languages. Pre-publication copy. Summer Institute of Linguistics. Veen, H. van der. 1929. Nota Betreffende de grenzen van de Sa1dansche taalgroep en bet aanverwante taalgebied. Tijdschrift voor Indische Taal-, Land- en Volkenkunde, 69:58-97. 132 TORAJA APPUDlCES A. Matrix "5 - South Sulawesi Stock - 63 wordlists NORTJ-IERN SOUTH SULAWESI FANlt,Y TORAJA 133 B. Hatru 6 - Central Sulawesi Stock - 16 vordlist8 BUNGKU SUBFAMILY ABOPADA 59 64 80 BETELEME MORl 62 68 79 92 MORlBAWAH (B) BAWAH 64 68 76 90 93 NUHA LANGUAGE 63 68 76 91 92 99 SOROAKO 59 62 77 82 80 86 87 KAWATAK 55 60 75 78 78 79 80 85 WAWANDOLA (G) PADOE 57 60 76 77 76 80 80 84 94 PABETA LANGUAGE 59 61 76 77 76 80 80 84 94 96 KAWAT 53 56 79 79 75 75 76 80 79 78 80 KOROPANSU 6 7 7 80 80 80 82 T 5 50 51 55 57 58 56 7 61 TOLAKl BPB TOLAIC 45 42 43 45 48 44 44 44 45 44 44 45 51 43 MANGKUTANA PAMONA 44 41 42 44 47 44 45 45 44 44 43 46 48 42 94 MALEKU (G) ..soUTH. SULAWE SJ MAP 134 TORAJA ell "p. of are•• urveyed Hap 1 - RorthctIIltral South Sul_.i Hap TORAJA 135 Hap 2 - Rortheastern South Sulawesi Hap 136 TORAJA